Thermoplastic Forming of Metallic Glass: Die Selection, Microforming and Property Characterisation

Thesis by

Amir Monfared

A Thesis in Fulfilment of the Requirements for the Degree of Doctor of Philosophy

School of Mechanical and Manufacturing Engineering

The University of New South Wales Sydney, Australia

August 2018

PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: Monfared

First name: Amir Other name/s: ;. Abbreviation for degree as given in the University calendar: PhD

School: Mechanical and Manufacturing Faculty: Engineering

Title: Thermoplastic Forming of Metallic Glass: Die Selection, Microforming and Property Characterisation

Abstract 350 words maximum: (PLEASE TYPE) Metallic glass (MG) is a promising class of materials with exceptional properties. Due to non-equilibrium amorphous structure, manufacturing of these alloys is challenging and might lead to structural and properties changes. Thermoplastic forming (TPF) is an efficient manufacturing technique of MGs. However, there are still many challenges (e.g. MG/die adhesion) and ambiguities such as the effect of TPF on apparent viscosity, mechanical property and structure which should be resolved. Accordingly, this thesis aims to examine TPF of MGs and improve the manufacturing of these alloys. The following lists the major findings of this thesis: 1) The dies with lower surface free energy (SFE) and higher bonding dissociation energy showed the least adhesion with MGs. Chemical adhesion and diffusionwere recognized as the primary adhesion mechanisms. New models were developed and verified for the evaluation of SFE of MGs and the work of adhesion between MGs and dies. 2) Apparent viscosity investigations revealed that at lower temperatures the apparent viscosity increased throughout the tests. However, at higher temperatures viscosity reached to a plateau. 3) The structural analyses of MGs revealed that TPF altered the diffractionpatterns size. The diameter size of the first ring of diffractionpattern increased afterTPF; yet the diameter size becomes larger with temperature. Due to the inverse relationship of interatomic spacing and ring size in diffractiontheory, it was concluded that interatomic spacing reduced after TPF confirming free volume annihilation. DSC analyses showed that the heat release of the sample thermoplastically formed is smaller than the as received MG; yet heat release decreases with increasing forming temperature. This verifies the occurrence of structural relaxation during TPF. 4) It was found that MGs becomes harder afterTPF; yet, the hardness increases with temperature rise. It was also demonstrated that dies played trivial role in the hardening of MGs. It was revealed that the load displacement curves of the samples after TPF exhibited less pop-ins compared with the as-received MGs. Based on viscosity variation, HRTEM analyses, DSC investigations and nanoindentation results, free volume annihilation via structural relaxation was identifiedas the primary hardening mechanism.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in DissertationAbstracts International (this is applicable to doctoral theses only) .

...... wJness Signature . . . 2.J/?/2-PloDate

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and re uire the a roval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

Acknowledgement

Firstly, I would like to express my sincere gratitude to my supervisor Professor

Liangchi Zhang for the continuous support of my Ph.D study and related research, his patience, encouragement, advice and immense knowledge. His guidance helped me in all the time of research and writing of this thesis.

I also express my deepest gratitude to my co-supervisor, Dr. Weidong Liu for his unselfish help and patience, guidance and crucial advice at different stages of the research, without whose assistance, this study would not have been successful.

I would like to thank all of the members of the Laboratory for Precision and Nano

Processing Technologies (LPNPT) for stimulating discussions and their assistance during experimentations in the past four years.

Last but not the least; I would like to thank my family: my parents and my brother for supporting me spiritually throughout writing this thesis and my life in general.

Abstract

Metallic glass (MG) is a promising class of material with exceptional properties. Due to non-equilibrium amorphous structure; manufacturing of these alloys is challenging and can lead to changes in structure and properties. Thermoplastic forming (TPF) is one of the most efficient manufacturing techniques of MGs. However, due to amorphous structure in the supercooled liquid region (SCLR), crystallisation and property changes are likely to occur in TPF. While some research has been conducted in this area, there are still many challenges and ambiguities (e.g. MG/die adhesion, apparent viscosity changes in TPF, effect of TPF on mechanical property and structure of MG) that should be addressed to obtain a technique with higher performance. Adhesion significantly affects the quality of manufacturing and net price of production through deteriorating the surface quality of MG and die.

Apparent viscosity plays a major role in TPF; however comprehensive research examining apparent viscosity variation in TPF is lacking. Apparent viscosity is important in revealing the deformation behaviour of MG in TPF. In addition, the effect of TPF on the structure and mechanical property of MGs remains unclear.

Given the importance of MG structure and property on the performance of MGs in service, exploring the structure and property of MGs after TPF is essential. Thus, this thesis aims to examine the TPF of MGs to address the challenges and ambiguities in this area and improve the efficiency of manufacturing of MGs.

The major findings of this thesis are listed as follow:

1) Several materials including electroless Ni-P, Si, SiC, Si3N4, alumina,

polytetrafluoroethylene (PTFE), sapphire and WC-Co were employed to examine

their adhesion status with La-based and Zr-based MGs. It was revealed that WC-

Co, sapphire and PTFE had the lowest adhesion status. Among these three

materials, WC-Co was identified as the best die due to its superior machinability,

excellent mechanical properties, higher working temperature and cheaper price.

2) Surface free energy (SFE) and bonding dissociation energy (BDE) of the dies

were identified as the parameters playing the major roles in the adhesion status.

The dies with lower SFE and higher BDE showed the least adhesion with MGs

and vice versa. Chemical adhesion and diffusion were recognised as the primary

adhesion mechanisms in TPF, as verified through energy dispersive spectroscopy

analyses and high resolution scanning electron microscope observations.

3) Due to the importance of SFE in TPF of MGs and the lack of appropriate model

in this area, a new model was developed in this thesis for the evaluation of SFE

of MGs and was verified with the available experimental data. In addition, a

novel model was introduced in this thesis for estimation of the work of adhesion

between MGs and dies and its validity was verified through the experimental

observations.

4) Given the importance of apparent viscosity in TPF, the apparent viscosity

variation of MG during TPF at different temperatures has been obtained in this

thesis. At lower temperatures where viscosity is far from equilibrium, the

apparent viscosity increased throughout the tests. However, at higher

temperatures where the apparent viscosity approaches equilibrium, viscosity

reaches a plateau with little variation. Viscosity variation demonstrates the free

volume annihilation in TPF.

5) The structure of MGs before and after TPF were thoroughly analysed through X-

ray diffraction and high resolution transmission electron microscope (HRTEM). iii

The analyses verified the amorphous structure of MG after TPF at all

temperatures. However it was revealed that TPF altered the diffraction patterns.

The diameter size of the first ring of diffraction pattern increased after TPF; yet

the diameter size became larger with temperature rise. Due to the inverse

relationship between interatomic spacing and ring size in diffraction theory, it

was concluded that interatomic spacing reduced after TPF, thus confirming free

volume annihilation through structural relaxation.

6) Differential scanning calorimetry (DSC) analyses of the as-received and

thermoplastically formed sample revealed the occurrence of structural relaxation.

By comparison with the as-received MG, heat release decreased for the sample

thermoplastically formed and decreased further with increasing the forming

temperature. The reduction of heat release verifies free volume annihilation.

7) TPF of MGs was conducted on two different dies, microchannelled and

microholed, at different temperatures. Microribs and microrods were

manufactured successfully without any adhesion. Based on the filling lengths of

the microribs and microrods, the apparent viscosity of the flow during the

experiments was calculated at each temperature. For microrods, the apparent

viscosity decreased from 6.6 ×1010 Pa.s at 450˚C (minimum allowable

temperature for microrod fabrication) to 1.6 ×107 Pa.s at 496˚C. It was

demonstrated that oxidation and capillary force played trivial roles in apparent

viscosity.

8) Mechanical properties are among the most attractive MG properties.

Accordingly, the effect of TPF on the mechanical properties of MG was

thoroughly examined under different conditions. It was found that MGs became

harder after TPF and the hardness increases with temperature rise. Further, iv

localised shear deformation as commonly observed around a Vickers indentation

mark of raw MGs does not occur in the MG after TPF. In addition, it was

demonstrated that dies played a trivial role in the hardness improvement of MGs.

9) It was revealed that the load displacement curves of the samples after TPF

exhibited less pop-ins compared with the as-received MGs. The curves become

even smoother with temperature rise. The analyses of the load displacement

curves at different loading rates proved the rate dependency of the pop-ins in

MGs. At higher loading rates, the load displacement curves had fewer pop-ins

and vice versa which is due to the activation of localised shear deformation in

MGs. At higher loading rates, a single localised shear is activated and grows

through MG. But when the loading rate is lower, the number of activated shear

bands becomes larger and the load displacement shows more pop-ins.

10) The hardening mechanism of MGs after TPF was revealed. Free volume

annihilation through structural relaxation was identified as the primary hardening

mechanism verified via viscosity variation, HRTEM analyses and

nanoindentation results.

Keywords:

Metallic glass; thermoplastic forming; adhesion; viscosity; mechanical properties; structural characterisation

List of nomenclature

A Area

c0 Constant related to Avogadro number

cf Free volume concentration

cf, eq Equilibrium free volume concentration

CS Surface area fractions

Die size

𝑑𝑑 D Fragility parameter

De Equivalent hydraulic diameter

fvacuum Degrees to which the surface atoms surrounded by vacuum g Perimeter h Oxidation layer thickness

kr Rate constant

First order rate constant

𝑘𝑘𝑐𝑐 Filling length

𝐿𝐿 Ldie Die length

P Pressure

P0 Pressure required for breaking the oxide layer

Q Volume flow rate

Re Reynolds number

S Formability factor t Time

ti Initial thickness

tf Final thickness

T Temperature

Tg Glass transition temperature

Tl Liquidus temperature of metallic glass

Tm Melting temperature

TX Crystallisation temperature

T* VFT temperature u Velocity distribution

V Molar volumes

W Width

Wad Work of adhesion x Reduced free volume

γ Surface free energy

vii

Geometrical overlap factor

𝛾𝛾0 Shear strain rate

𝛾𝛾̇ Surface enthalpy 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 ∆𝐻𝐻 Interface enthalpy 𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 ∆𝐻𝐻 Strain rate

𝜀𝜀̇ η Viscosity

η0 Pre-exponential factor

Apparent viscosity

𝜂𝜂𝑎𝑎𝑎𝑎𝑎𝑎 ηeq Equilibrium state of viscosity

Wetting angle

𝜃𝜃 Poisson ratio

𝜇𝜇 Velocity

𝜐𝜐 υ* Critical free volume concentration

Average free volume per atom

𝜐𝜐𝑓𝑓 ρ Density

σflow Flow stress

τ Viscous stress

viii

List of Abbreviation

BDE Bonding dissociation energy

BMG/MG Bulk metallic glass/Metallic glass

CSM Cooperative shear model

CV Controlled volume

DSC Differential scanning calorimetry

EDS Energy dispersive spectroscopy

FIB Focused ion beam

GFA Glass forming ability

HRTEM High resolution transmission electron microscope

HV Vickers hardness

MEMS Micro-electro-mechanical system

MFC Micro fuel cell

PCDSE Polycrystalline diamond square end mill

PEL Potential energy landscape

PMMA Poly(methyl methacrylate)

PTFE Polytetrafluoroethylene

SCLR Supercooled liquid region

SEM Scanning electron microscope

SFE Surface free energy

STZ Shear transformation zone

TPF Thermoplastic forming ix

XRD X-ray diffraction

TTT Time-temperature-transformation

UNSW University of New South Wales

VFT Volger-Fulcher-Tammann

List of Figures

Figure 1.1 Flow chart of the different sections of the thesis ...... 7

Figure 2.1 Critical cooling rate for Ribbon MG, BMG and silicate glass [3]...... 12

Figure 2.2 Schematic potential energy landscape for BMG [29]...... 13

Figure 2.3 Icosahedral structure of MGs a) 2D b) 3D [25] ...... 15

Figure 2.4 a) Tensile strength with Young’s modulus and b) Vickers hardness with

Young’s modulus for different materials [1] ...... 19

Figure 2.5 Cooling rate cycle for different manufacturing processes of amorphous alloys [7]...... 24

Figure 2.6 Schematic of TPF of MGs ...... 29

Figure 2.7 Nanorod of Pt-based MG [62] ...... 30

Figure 2.8 Temperature dependent crystallisation time and viscosity for a Zr-based

BMG [7] ...... 32

Figure 2.9 Atomic jump in free volume model [44] ...... 41

Figure 2.10 Movement of a STZ during applying shear stress [108] ...... 44

Figure 2.11 Viscosity variation with temperature for a number of amorphous materials [1] ...... 47

Figure 3.1 DSC result of La60.5Al16.3 (Cu, Ni)23.2 ...... 54

Figure 3.2 XRD results of La60.5Al16.3 (Cu, Ni)23.2...... 54

Figure 3.3 XRD result of Zr58.5 Cu15.6 Ni12.8 Al10.3 Nb2.8 ...... 55

Figure 3.4 Toshiba precision glass moulding machine (GMP-211) ...... 57

Figure 3.5 (a) SEM image of the electroless Ni-P die surface after TPF, and (b) the

EDS result of the residual materials on the die surface after demoulding ...... 58

Figure 3.6 (a) Optical microscope image of the La-based MG and Si die, and (b) high resolution SEM image of the interface layer...... 59

Figure 3.7 (a) EDS mapping for the La-based MG, interface and Si after TPF, (b)

EDS point analysis of the interface layer between the La-based MG and Si ...... 59

Figure 3.8 SEM image of the La-based MG after TPF on a PTFE die ...... 60

Figure 3.9 Optical microscope images of the SiC die surface after TPF with (a) the

La-based MG, and (b) the Zr-based MG ...... 60

Figure 3.10 Optical microscope image of the Si3N4 die after TPF with the La-based

MG ...... 61

Figure 3.11 Optical microscope image of the alumina die after TPF with the La- based MG ...... 62

Figure 3.12 Optical microscope image of the sapphire die after TPF with (a) the La- based MG, and (b) the Zr-based MG ...... 63

Figure 3.13 Optical microscope image of the WC-Co die after TPF with Zr-based

MG ...... 64

Figure 3.14 BDE of the constituent bonds of the die materials selected for investigation ...... 67

Figure 3.15 BDE of (a) C-C, Zr-C and La-C, and (b) Si-C, La-C and Zr-C (c) Al-O,

Zr-O and La-O (d) Si-N, La-N and Zr-N ...... 70

Figure 4.1 Zygo instrument in nano and precision engineering lab at UNSW ...... 82

Figure 4.2 Load displacement curve of the sample thermoplastically formed at

450˚C...... 83

Figure 4.3 Load displacement curve of the sample thermoplastically formed at 460˚C

...... 84

xii

Figure 4.4 Load displacement curve of the sample thermoplastically formed at 470˚C

...... 84

Figure 4.5 Load displacement curve of the sample thermoplastically formed at 480˚C

...... 85

Figure 4.6 Load displacement curve of the sample thermoplastically formed at 490˚C

...... 85

Figure 4.7 Load displacement curve of the sample thermoplastically formed at 496˚C

...... 86

Figure. 4.8 Maximum strains at different temperatures in TPF ...... 88

Figure 4.9 Viscosity variations with time at 450˚C ...... 90

Figure 4.10 Viscosity change with time at 460˚C ...... 91

Figure 4.11 Viscosity change with time at 470˚C ...... 91

Figure 4.12 Viscosity change with time at 480˚C ...... 92

Figure 4.13 Viscosity change with time at 490˚C ...... 92

Figure 4.14 Viscosity change with time at 496˚C ...... 93

Figure 4.15 Variation of viscosity with time of Pd41Ni10Cu29P20 at 555 K and 15 MPa due to structural relaxation [101] ...... 94

Figure 4.16 XRD result of the sample thermoplastically formed at 490˚C ...... 96

Figure 4.17 XRD result of the sample thermoplastically at 496˚C ...... 96

Figure 4.18 SEM image of the as received sample prepared by FIB for TEM analysis

...... 97

Figure 4.19 SEM image of the thermoplastically formed sample at 450 ˚C sample prepared by FIB for TEM analysis...... 97

Figure 4.20 SEM image of the thermoplastically formed sample at 496 ˚C prepared by FIB for TEM analysis ...... 98 xiii

Figure 4.21 Structure and diffraction pattern of the as-received material ...... 99

Figure 4.22 Structure and diffraction pattern of the sample thermoplastically formed at 450˚C ...... 100

Figure 4.23 Structure and diffraction pattern of the sample thermoplastically formed at 496 ˚C ...... 100

Figure 4.24 DSC result of the as-received MG ...... 101

Figure 4.25 DSC result of the thermoplastically formed MG at 450˚C ...... 102

Figure 4.26 DSC result of the thermoplastically formed MG at 496˚C ...... 102

Figure 5.1 Schematic and specifications of PCDSE ...... 106

Figure 5.2 Nano 350 FG 5 Axis Ultra Precision Lathe ...... 107

Figure 5.3 Microchannelled WC-Co die ...... 107

Figure 5.4 Fabricated WC-Co microchannelled at higher magnification ...... 108

Figure 5.5 Zygo image of the channel profile on the WC-Co ...... 108

Figure 5.6 Roughness of the channel surface...... 109

Figure 5.7 Microrib thermoplastically formed at 450˚C ...... 110

Figure 5.8 Microrib thermoplastically formed at 460˚C ...... 111

Figure 5.9 Microrib thermoplastically formed at 470˚C ...... 111

Figure 5.10 Microrib thermoplastically formed at 480˚C ...... 112

Figure 5.11 Microrib thermoplastically formed at 490˚C ...... 112

Figure 5.12 Microrib thermoplastically formed at 496˚C ...... 113

Figure 5.13 The schematic tube for force analysis ...... 115

Figure 5.14 Filling length variations with temperature for microribs ...... 117

Figure 5.15 Apparent viscosity variations with temperature for TPF of LM106a using a microchannelled die ...... 118

Figure 6.1 Diamond tool employed for manufacturing of the microholed die ...... 123 xiv

Figure 6.2 Diamond tool employed for manufacturing of the microholed die ...... 123

Figure 6.3 Microholed WC-Co die ...... 124

Figure 6.4 Microrod fabricated at 496 ˚C by microforming ...... 126

Figure 6.5 Nikon DS-Ri2 optical microscope ...... 127

Figure 6.6 Optical microscope analysis of the side view of the microrods fabricated at 490˚C ...... 127

Figure 6.7 Optical microscope analysis of the side view of the microrods fabricated at 496˚C ...... 128

Figure 6.8 Optical microscope analysis of the side view microrods fabricated at

496˚C at higher magnification ...... 128

Figure 6.9 Top view of the microrod fabricated at 496˚C ...... 129

Figure 6.10 Surface roughness of the microrod fabricated at 496 ˚C ...... 130

Figure 6.11 Microrod length variations with temperature ...... 131

Figure 6.12 Apparent viscosity variation at different temperature in TPF ...... 132

Figure 6.13 HRTEM analysis of the as-received MG...... 136

Figure 6.14 HRTEM analysis of the microrod fabricated at 450˚C ...... 137

Figure 6.15 HRTEM analysis of the microrod fabricated at 496 ˚C ...... 137

Figure 7.1 Struers Vickers Microhardness tester (Dura scan-80 G5) ...... 142

Figure 7.2 Hysitron nanoindentation instrument ...... 142

Figure 7.3 Hardness of the as-received materials and samples thermoplastically formed at different temperatures on flat die ...... 143

Figure 7.4 Optical microscope image of the indentation mark at a) load = 1 kgf and b) load = 5 kgf on the as-received material ...... 145

Figure 7.5 Optical microscope images of the indentation marks under (a) 1kgf, (b) 2 kgf (c) 5 kgf and (d) 10 kgf on the sample formed at 450˚C ...... 146 xv

Figure 7.6 Hardness of the microribs fabricated at different temperatures...... 148

Figure 7.7 Hardness of the fabricated microrods at different temperatures ...... 149

Figure 7.8 Load displacement curves of the a) as received and thermoplastically formed samples at b) 460˚C c) 480˚C d) 496˚C under the loading rate of 2×10-2 1/s

...... 151

Figure 7.9 Load displacement curve of LM106a at loading rate of a) 2×10-1 1/s b)

5×10-2 1/s c) 2×10-2 1/s d) 5×10-3 1/s ...... 152

Figure 7.10 Load displacement curve of the sample thermoplastically formed at

460˚C and loading rate of a) of 2×10-1 1/s b) 5×10-2 1/s c) 2×10-2 1/s d) 5×10-3 1/s 153

Figure 7.11 Load displacement curve of the sample thermoplastically formed at

480˚C at loading rate of a) 2×10-1 1/s b) 5×10-2 1/s c) 2×10-2 1/s d) 5×10-3 1/s ...... 154

Figure 7.12 Load displacement curve of the sample thermoplastically formed at

496˚C at loading rate of a) 2 × 10-1 1/s b) 5 × 10-2 1/s c) 2 × 10-2 1/s d) 5 × 10-3 1/s 155

Figure 7.13 Logarithmic variation of viscosity of LM106a with time at different forming temperatures ...... 157

Figure 7.14 DSC results of the as-received and thermoplastically formed MGs at

450˚C and 496˚C ...... 159

xvi

List of Tables

Table 2.1 The developing years of different MG systems [2] ...... 11

Table 2.2 Properties and application field of MG [5] ...... 23

Table 2.3 Properties of manufacturing techniques of different dies in TPF [6] ...... 35

Table 3.1 Summary of adhesion status and the SFE of MGs and die materials ...... 66

Table 3.2 SFE of MGs based on the presented model ...... 73

Table 3.3 Contribution of capillary pressure on TPF of two different MGs on quartz

...... 74

Table 3.4 Effect of material and size on capillary pressure in TPF of MGs ...... 75

Table 3.5 SFE of MGs and estimated work of adhesion in different circumstances . 77

Table 4.1 Calculated viscosity of LM106a at different temperatures ...... 88

xvii

TABLE OF CONTENTS

Acknowledgement ...... i

Abstract ...... ii

List of nomenclature ...... vi

List of Abbreviation ...... ix

List of Figures ...... xi

List of Tables ...... xvii

Chapter 1 Introduction ...... 1

1.1 Motivation ...... 2

1.2 Present state of the problem ...... 3

1.3 Objective and scope of the thesis ...... 4

1.4 Thesis structure ...... 5

Chapter 2 Literature Review ...... 8

2.1 Metallic glass (MG) ...... 9

2.1.1 Development of MGs...... 9

2.1.2 Amorphous structure of MGs...... 12

2.1.3 Properties of MGs ...... 16

2.1.3.1 Physical properties ...... 16

2.1.3.2 Chemical properties ...... 18

2.1.3.3 Mechanical properties ...... 18

xviii

2.1.4 Applications of MGs ...... 21

2.1.4.1 Biomedical applications ...... 21

2.1.4.2 Nanotechnology and MEMS ...... 22

2.2 Conventional manufacturing techniques of MGs ...... 23

2.2.1 Casting ...... 24

2.2.2 Machining ...... 26

2.3 Thermoplastic forming (TPF) ...... 27

2.3.1 Procedure ...... 28

2.3.3 Challenges of TPF of MGs ...... 31

2.4 Mechanical behaviour of MG in TPF ...... 36

2.4.1 Inhomogeneous and homogeneous deformations of MGs ...... 37

2.4.1.1 Inhomogeneous deformation ...... 37

2.4.1.2 Homogeneous deformation ...... 39

2.4.2 Theoretical model of plastic deformation ...... 40

2.4.2.1 Free volume model ...... 40

2.4.2.2 STZ model ...... 43

2.4.3 Viscosity ...... 44

2.4.3.1 Temperature effect ...... 46

2.4.3.2 Strain rate effect ...... 47

2.4.3.3 Apparent viscosity in TPF ...... 48

2.5 Summary ...... 49

xix

Chapter 3 Die Selection ...... 52

3.1 Materials selection ...... 53

3.2 Experimental set up ...... 56

3.3 Adhesion status of different dies and MGs ...... 57

3.3.1 Die made of electroless Ni-P ...... 57

3.3.2 Die made of the Si wafer ...... 58

3.3.3 The PTFE die ...... 59

3.3.4 SiC ...... 60

3.3.5 Si3N4...... 61

3.3.6 Alumina ...... 61

3.3.7 Sapphire ...... 62

3.3.8 WC-Co ...... 63

3.4 Adhesion mechanism of dies and MG ...... 64

3.5 SFE of MGs ...... 71

3.5.1 A model for calculation of SFE of MG ...... 71

3.5.2 SFE in TPF...... 73

3.6 Work of adhesion between MGs and die materials ...... 75

3.7 Summary ...... 77

Chapter 4 TPF of MG on Flat Die ...... 79

4.1 Experimental procedure ...... 80

4.2 Characterisation after TPF ...... 83

4.2.1 Load displacement curves ...... 83

4.2.2 Formability...... 87

4.2.3 Apparent viscosity ...... 88

4.2.4 XRD analysis ...... 95

4.2.5 HRTEM analysis...... 96

4.2.6 DSC analyses ...... 100

4.3 Summary ...... 103

Chapter 5 Micro Forming of MG Using a Microchannelled Die...... 104

5.1 Fabrication and characterisation of die...... 105

5.2 Experimental set up ...... 109

5.3 Characterisation of the microribs ...... 110

5.3.1 Microscopy analysis ...... 110

5.3.2 Roughness analysis ...... 113

5.3.3 Apparent viscosity ...... 113

5.4 Summary ...... 120

Chapter 6 Micro Forming of MG Using a Microholed Die ...... 121

6.1 Fabrication and characterisation of the die ...... 122

6.2 Experimental set up ...... 124

6.3 Characterisation after TPF ...... 125

6.3.1 Microscopy analysis ...... 126

6.3.2 Roughness analysis ...... 129

xxi

6.3.3 Apparent viscosity ...... 130

6.3.4 Structural analysis ...... 135

6.4 Summary ...... 137

Chapter 7 Mechanical Property Characterisation of MG after TPF ...... 139

7.1 Experimental procedure ...... 141

7.2 Microhardness results ...... 142

7.2.1 Flat die ...... 143

7.2.2 Microchannelled die ...... 147

7.2.3 Microholed die ...... 148

7.3 Nanoindentation results ...... 150

7.3.1 The effect of TPF on load-displacement...... 150

7.3.2 Loading rate effect on load-displacement curve ...... 152

7.3.2.1 As-received MG ...... 152

7.3.2.2 Thermoplastically formed MG at 460˚C ...... 153

7.3.2.3 Thermoplastically formed sample at 480 ˚C ...... 153

7.3.2.4 Thermoplastically formed sample at 496˚C ...... 154

7.4 Mechanism of mechanical property improvement ...... 155

7.5 Summary ...... 163

Chapter 8 Conclusions and Future Research ...... 165

8.1 Conclusions ...... 166

8.2 Future research ...... 169

xxii

List of Publications ...... 171

Journal papers ...... 171

Conference ...... 171

References ...... 173

xxiii

Introduction

Chapter 1 Introduction

Introduction

1.1 Motivation

Metallic glasses (MGs) are new alloys that have attracted a great deal of attention in the past few decades [1-3]. Due to their amorphous structure, these alloys possess unique properties such as high elastic limit, strength, hardness and wear resistance

[1, 4]. According to these characteristics, they have a wide range of applications in different fields such as biomedicine, transformers, sports industries and fuel cells [1,

5].

Due to their brittleness, higher viscosity compared with polymers and metastable structure, manufacturing has been a key challenge in MG field [6, 7]. Casting was the first and oldest manufacturing technique utilised for MGs [6]. However, this method is not able to fabricate complex shaped products particularly in micro and nano scales [6]. Machining of these alloys would be extremely difficult due to their brittleness and the risk of temperature increase and structural change during the process [8, 9]. However the unique superplasticity behaviour of MGs within the supercooled liquid region (SCLR) provides the opportunity of precision moulding via thermoplastic forming (TPF) [6, 7]. MGs can be shaped to nanometer scales with excellent roughness through this technique [10]. However, TPF encounters some challenges and ambiguities that must be resolved to increase its productivity. The motivation of this thesis is to provide an in depth understanding of the behaviour of

MGs in TPF.

Introduction

1.2 Present state of the problem

MGs are promising class of alloy with exceptional properties such as high strength, high elastic limit, and high wear and corrosion resistance. However the unstable amorphous structure of these alloys creates a major challenge in their manufacturing.

In the past two decades TPF has been utilised as a manufacturing technique of MGs and a great deal of research has been conducted to find an efficient technique for fabricating MG products. However there are some challenges and ambiguities involved in using TPF to manufacture MGs that must be resolved.

MG/die adhesion in TPF is a pivotal issue because it deteriorates the surface quality of the die and/or the MG and increases the net price of manufacturing the products.

There is no research providing an in depth investigations of the MG/die adhesion and its mechanism in TPF. In addition, no model has been created to estimate the

MG/die work of adhesion in TPF.

Apparent viscosity of MGs is a primary factor in TPF and might change during the

TPF process due to crystallisation. Crystallisation reduces the formability of MG and increases the apparent viscosity. Temperature is one of the most important parameter affecting the apparent viscosity in TPF. However, a systematic research that considers the effect of temperature on apparent viscosity in TPF is lacking.

MGs possess an unstable amorphous structure that tends to be crystallised. In TPF,

MGs are held and shaped at high temperatures. This increases the risk of structural change and/or crystallisation. Changing structure may alter the property of MGs and affect the performance of these alloys in service. Considering the application of MGs 3

Introduction

in highly sensitive areas such as biomedicine and aerospace, this issue is highly

critical. However, limited data are available on the effect of TPF on the structure of

MGs.

MGs are recognised as alloys with exceptional mechanical properties that make them

highly appropriate candidates for different applications. When applying high

temperatures, it is likely that the mechanical properties of MGs are affected by TPF

through different phenomena. However, thus far no any systematic research has

investigated the mechanical property changes of MGs during TPF.

1.3 Objective and scope of the thesis

This thesis provides in-depth understanding of MGs’ behaviours in TPF with the aim

of creating understanding in relation to manufacture MGs products with better

qualities. The primary objectives of this thesis are as follows:

I. Explore the adhesion status of different materials with MGs and identify the best

die materials in TPF of MGs to create products with a superior surface quality.

II. Investigate the primary parameters affecting the adhesion status of dies with

MGs and reveal MG/die adhesion mechanisms in TPF

III. Develop a model for estimation of surface free energy (SFE) of MGs and work

of adhesion between dies and MGs.

IV. Explore apparent viscosity variation of MGs and fluid behaviour in TPF in

different conditions

V. Explore the effect of TPF at different temperatures on the structure and

diffraction pattern of MGs

Introduction

VI. Explore the effect of TPF parameters such as temperature and die on MGs’

mechanical properties such as hardness and load displacement curves

1.4 Thesis structure

This thesis is presented in eight Chapters. Chapter 1 introduces the research

motivation, the research gaps, the research objectives and outline of the thesis.

Chapter 2 presents a comprehensive literature review on the research progress in

MGs. The history, structure, properties and applications of MGs are discussed to

introduce the significance of these alloys. Different conventional manufacturing

techniques including casting and machining along with their limitations are also

presented. This Chapter also presents a detailed analysis of the TPF of MGs as the

key concern of this thesis and discusses its applications and challenges. The most

common models of plastic deformation and the mechanical properties of MGs are

also examined. In addition, given the importance of viscosity in TPF, the effects of

different parameters on viscosity are discussed.

The primary objective of Chapter 3 is to explore the best die material to use in TPF

of MGs. Several materials with a wide range of chemical, physical and mechanical

properties are selected for analyses. After TPF, the surface of each die is

characterised by microscopic and elemental analyses and the dies are listed

according to their adhesion status. The key parameters affecting the MGs/die

adhesion are examined and the mechanism of adhesion is revealed. A novel model is

Introduction

developed to calculate the work of adhesion between MGs and dies in TPF. The selected die is later used to conduct the TPF of MGs experiments.

After die selection, Chapters 4, 5 and 6 examine the effects of TPF parameters on the fluid behaviour, structure and properties of MGs. In Chapter 4, TPF of MGs is carried out on a flat die at different temperatures. The stress-strain curve, formability and apparent viscosity variation with time at each temperature are calculated. The structure of MGs after TPF is characterised by using X-ray diffraction (XRD) and high resolution transmission electron microscope (HRTEM) before and after TPF. In

Chapters 5 and 6, TPF of MGs at different temperatures is conducted by using microchannelled and microholed dies, respectively. The apparent viscosity at each temperature is calculated according to the filling length of the samples and the effects of the primary parameters such as oxidation and capillary pressure on apparent viscosity are estimated. The structures of the microrods are also characterised using HRTEM. Chapter 7 examines the mechanical properties of the samples before and after TPF by conducting microhardness and nanoindentation tests on the thermoplastically formed samples obtained in Chapters 4, 5 and 6.

Considering the results of mechanical and structural analyses of MGs, the mechanism of property enhancement from using TPF is revealed. Chapter 8 presents the conclusions of the research and discusses avenues for future research. Fig 1.1 presents the outline of the thesis Chapters.

Introduction

Figure 1.1 Flow chart of the different sections of the thesis

Literature Review

Chapter 2 Literature Review

Literature Review

2.1 Metallic glass (MG)

2.1.1 Development of MGs

Glass is a class of materials that has been known and utilized in daily life for several thousand years [11]. However, other types of glass including polymer glass and MGs were unknown until several decades ago [3]. The first MG with chemical composition Au75Si25 were introduced by a group at Caltech, USA in 1960 [12]. This family is distinguished from other sorts of glasses owing to metallic bonds among their atoms. MGs possess a combination of characteristics of metal and glass due to metallic bonds and amorphous structure. This behavior makes them attractive for a number of applications such as nano and micro manufacturing, microelectromechanical-system (MEMS), fuel cells and sports industries [13-15].

The development of MGs can be categorized to two stages, the ribbon and the bulk.

The MG ribbons with thicknesses of lower than 100 µm were primarily based on noble metals, such as Au-Si, Pd-Ni-Si and Pd-Si-Ag [16-18]. Casting was the first method of fabrication of ribbon amorphous alloys. Rapid solidification technique such as splat quenching was utilized for casting of MGs to inhibit crystallization.

This method restricted the thickness of the products to less than 100 µm and was only suitable for ribbons.

Meanwhile a series of attempts [16-18] were made to fabricate BMGs with a higher glass forming ability (GFA). In 1974, Chen [19] prepared the first BMGs from Pd-

Cu-Si using a simple suction-casting method at a significant cooling rate of 103 K/s.

Literature Review

Turnbull [20] predicted that the ratio of Tg (glass transition temperature) to Tm

(melting point) could be used as a criterion for making MG components with high

GFA. Accordingly the MGs with Tg/Tm > 2/3 possessed very high GFA and could only crystallize within a very narrow temperature range. Based on this, Turnbull et al. [21, 22] successfully made Pd-Ni-P MGs at a cooling rate of 10 K/s. Although it was important, this discovery did not revolutionize the field due to the cost of noble metals.

The real breakthrough was after discovery of multicomponent MGs in 1990s at

Tohoku University. Inoue et al. [13, 23] found excellent GFA in rare-earth-based- alloys, and made MGs of La-Al-Ni and La-Al-Cu using water cooling Cu molds at the cooling rate of 1 K/s. They also developed similar quaternary and quinary MGs

(e.g., La–Al–Cu–Ni and La–Al–Cu–Ni–Co BMGs) at the cooling rates below 100

K/s. They could successfully reach casting thickness of several centimeters [24].

Meanwhile, some transition-based MGs (e.g. Zr-Al-Cu-Ni) with high GFA were developed in 1990s at California institute of technology named Vitreloy family [25].

Afterward several MGs systems including Mg-Cu-Y,-Zr-Al-Ni-Cu, Zr-Al-Ni-Cu(Ti,

Nb), Zr-Ti-Cu-Ni-Be, Ti-Ni-Cu-Sn, Cu-Zr-Ti-Ni, Nd-Fe-Co-Al, Fe-Co-Ni-Zr-Nb-B,

Fe-Al-Ga-P-C-B, Pr-Cu-Ni-Al and Pd-Ni-Cu-P were gradually developed. In Table

2.1 a list of different BMG systems and their developing years has been tabulated.

Literature Review

Table 2 .1 The developing years of different MG systems [2]

BMG systems Year Pd-Cu-Si 1974 Pt-Ni-P 1975 Au-Si-Ge 1975 Pd-Ni-P 1982 Mg-Ln-Cu (Ln=lanthanide metal) 1988 Ln-Al-TM(TM=group transition 1989 metal) Zr-Ti-Al-TM 1990 Ti-Zr-TM 1993 Zr-Ti-Cu-Ni-Be 1993 Nd(Pr)-Al-Fe-Co 1994 Zr-(Nb,Pd)-Al-Fe-Co 1995 Cu-Zr-Ni-TI 1995 Fe-(Nb,Mo)-(Al,Ga)-(P,C,B,Si,Ge) 1995 Pd-Cu(Fe)-Ni-P 1996 Co-(Al,Ga)-(P,B,Si) 1996 Fe-(Zr,Hf,Nb)-B 1996 Co-Fe-(Zr,Hf,Nb)-B 1996 Ni-(Zr,Hf,Nb)- (Cr,Mo)-B 1996 Ti-Ni-Cu-Sn 1998 La-Al-Ni-Cu-Co 1998 Ni-Nb 1999 Ni-(Nb,Cr,Mo)-(P,B) 1999 Zr-based glassy composites 1999 Zr-Nb-Cu-Fe-Be 2000 Fe-Mn-Mo-Cr-C-B 2002 Ni-Nb-(Sn,Ti) 2003 Pr(Nd)-(Cu,Ni)-Al 2003

Figure 2.1 schematically shows the critical cooling rate for establishing ribbon MGs,

BMGs, and silicon glass. It is evident the ribbon MGs possess the highest and silicate glass has the lowest critical cooling rate.

Literature Review

Figure 2.1 Critical cooling rate for Ribbon MG, BMG and silicate glass [3]

It should be mentioned that in spite of the progress in fabrication of MGs with high

GFA and lower critical cooling rate, the method is still quite empirical. There are some criteria such as, higher number of elements, greater atomic size difference and large negative heat of mixing which might help predict GFA [26]. However, a model that correlates all the parameters and accurately predicts GFA of different combination systems is still lacking [26].

2.1.2 Amorphous structure of MGs

From energy point of view, the crystalline phase is the equilibrium state and supercooled liquid with amorphous structure is non-equilibrium state when a melt is cooled down from melting point. MGs are essentially supercooled liquid with amorphous structure. The question is why in spite of non-equilibrium state, amorphous structure is achievable at room temperature.

Literature Review

When a melt is cooled down across its melting temperature, crystallization cannot occur immediately, though the crystal structure has a lower free energy than the liquid. This is mainly because of the effect of SFE [27]. The melt thus gets into a deep liquid state, called supercooled liquid. If the cooling rate is sufficiently high, the atomic rearrangement in the supercooled liquid towards crystallization would become increasingly difficult, owing to the significant increase of viscosity [3, 28] .

The critical cooling rate to freeze liquid configuration is in the range of 105 to 106

K/s. This phenomenon can be described by potential energy landscape (PEL) model.

A schematic of PEL for a Pd-based MG is illustrated in Fig. 2.2.

Figure 2.2 Schematic potential energy landscape for BMG [29]

The local atomic clusters fluctuate among the metabasins (smaller basins) of the

PEL. With decreasing the temperature, the local clusters would trap into the metabasins, which further induces the increase in viscosity and decreases the chance of climbing out of the metabasin. The supercooled liquid can eventually be vitrified near one particular temperature called the glass transition temperature Tg, below

Literature Review

which the atomic clusters are frozen into a solid state called MGs [3]. Accordingly

MGs contain high degree of dense randomly packed atomic configurations. The random structure causes the densities of MGs to be around 0.3 to 1% lower than fully crystallized counterparts [1, 2].

The key factor in fabricating MGs is suppressing crystalline particles in the structure and due to this, kinetic of crystal formation has been considered as pivotal controlling factor in MGs. Good glass formers have the near composition of eutectic liquids when a liquid transforms to two identical solids. Alloys based on noble elements such as Pd-based and Au-based systems require very low cooling rate of around 10 K/s to inhibit crystallisation. In these systems BMGs with thickness of around 10 mm can be prepared with low cooling rate.

Crystal nucleation rate is significantly influenced by the diffusivity of the elements which itself is a function of liquid viscosity. Dense liquids with more stability have higher GFA. Local icosahedral order has been considered as the most probable atomic motif of stable supercooled liquids and MGs. In this system crystallites are less prone to grow compared with BCC, HCP or FCC systems owing to their highly close packed structure and lack of translational periodicity. The schematic feature of densely packed atoms in icosahedral structure has been shown in Fig. 2.3 [25].

Literature Review

Figure 2.3 Icosahedral structure of MGs a) 2D b) 3D [25]

Atomic size ratio between solute and solvent atoms in an MG system takes the key part in atomic packing efficiency. A ratio close to 0.902 is considered as the highest atomic packing in icosahedra-like cluster [25].

Glass structures possess short to medium range orderings and lack any long range ordering [30]. Short range generally means a nanoscale local zone consisted of a core atom and its neighbors. Short range ordering of MGs can be revealed in TEM analysis and diffraction pattern [31, 32]. In contrast to the ordered structure of crystalline materials, the structure is lacking any order in TEM images of MGs [31,

32]. Diffraction pattern of MGs has also a halo shape pattern without any clear spots or rings [31, 32].

According to PEL, when a glass material is supercooled from its molten state, the atoms trap into a series of basins [29]. In these basins, atoms have more opportunity to form stable local clusters. However, out of the basins, the atomic arrangement would rather be random or disordered. The random atomic arrangement causes MGs

Literature Review

to have two distinct formability behaviors depending on temperature. At SCLR, the thermal vibration of atoms is sufficient to crossover of configurations over the whole energy landscape, leading to the superplasticity of MGs. At lower temperatures, however, atomic rearrangements are biased towards the lowest saddle, leading to heterogeneous deformation and shear banding failure [33, 34].

The random atomic arrangement of MGs amorphous structure causes these alloys to have extraordinary characteristics. The most notable properties of MGs will be discussed in the next section.

2.1.3 Properties of MGs

In this section some of the most prominent physical, chemical and mechanical properties of MGs will be discussed.

2.1.3.1 Physical properties

Physical properties are essential in better understanding of MGs behaviours in TPF.

Density, specific heat, thermal expansion and diffusivity are among the prominent physical properties.

i. Density:

It is known that the density of MGs is lower than the crystalline counterparts.

Density of BMG is approximately 0.3% to 1% lower than the crystalline

counterparts. This difference becomes even larger and reaches to 2% to 3%

between ribbons and crystalline counterparts [1]. BMGs have higher density than

Literature Review

the ribbon due to their closed packed structure. The difference in density of

BMGs and crystalline counterparts originates from the randomly packed structure of the former ones [1].

ii. Thermal expansion:

Coefficient of thermal expansion is a critical parameter in net processing of

MGs via TPF as a result of its effect on dimensional accuracy and thermal

stresses [35]. Large difference between coefficients of thermal expansion of die

and MGs leads to thermal stresses in MG products. Generally the coefficients of

thermal expansion of MGs are larger than the crystallized counterpart. Even in

MGs the coefficient of thermal expansion in SCLR is much larger than the

glassy state. This rapid change near Tg is attributed to the activation of new

mode of structural relaxation in the SCLR [1].

iii. Diffusion:

Diffusion is a physical property that takes a major part in the behaviour of MGs,

particularly in SCLR. In crystalline metals diffusion is a simple atom jump

phenomenon; however, in MGs diffusion is not necessarily a simple-jump atom

model. Based on the experimental results, it is found that simple jump and

cooperative motion of atoms would be the diffusion mechanisms in glassy state

and SCLR, respectively which causes the diffusivity of the latter state would be

much faster than the former one [1, 36].

Literature Review

2.1.3.2 Chemical properties

Chemical properties of MGs are particularly important when the working environment is harsh like in physiological medium, high temperatures and oxidising atmosphere. Generally MGs possess lower tendency of corrosion, owing to lack of defects such as grain boundary, dislocations and twins in their structure. These defects are in higher energy states compared with the bulk materials and are among the least resistive regions in front of corrosion and oxidising. However, given the lack of such high energy state areas, the corrosion resistance of MGs would be higher. In addition, high cooling rate during processing results in high homogeneity and further enhances the corrosion resistance of MGs [1, 37-39].

2.1.3.3 Mechanical properties

Mechanical properties play the major role in manufacturing and applications of

MGs. Here the notable mechanical properties of MGs are discussed.

One of the most eminent mechanical properties of MGs is their high strength and hardness which make them ideal candidates for a wide range of applications [25].

Strength approaching the theoretical value is achievable for these alloys. Tensile fracture strength and Vickers hardness have linear relationship with elastic modulus in amorphous and crystalline alloys. The slope of tensile fracture strength or Vickers hardness with elastic modulus indicates the elastic limit of materials. As shown in

Fig. 2.4, this slope is steeper for BMGs compared with crystalline alloys. This indicates that BMGs possess higher elastic limit compared with crystalline alloys which is attributed to high homogeneity over the whole composition range [1]. 18

Literature Review

Elastic deformation as much as 2% permits large-scale reversible deflection and elastic energy storage (resilience) for applications in sporting goods and springs [40].

Figure 2.4 a) Tensile strength with Young’s modulus and b) Vickers hardness with Young’s modulus for different materials [1]

Superior specific strength, large ductility in bending, low coefficient of friction and high wear resistance are among other unique mechanical properties of BMGs [4, 40].

MGs behave differently in tension and compression [41, 42]. While the stress–strain curves show no plasticity under tensile loading, a small amount of plastic strain under compressive loading is observed which is due to the extra constraint provided

Literature Review

in this configuration. In tension, deformation typically occurs through a single shear band and material fails by shear rupture of this band with very little plastic strain [41,

43]. In the absence of dislocation-mediated crystallographic slip, which is the typical mode of deformation for crystalline alloys, MGs undergo highly heterogeneous deformation by formation of localised shear bands [40, 44]. On the other hand, in compression, shear bands can accommodate very large localised plastic strains, giving MGs ductility. BMGs exhibit elastic-perfectly plastic deformation behavior under compressive loading at room temperature. The plastic stress–strain response displaying serrations, which correspond to the formation of shear bands [43]. BMGs are also susceptible to cyclic loading. Shear localization occurs during cyclic loading are responsible of damage under this mode of loading [40, 45].

A number of techniques have been utilized for improving the mechanical properties of MGs. It is reported that formation of nanocrystallites in MGs has improved the mechanical properties of MGs. Such in-situ formed nanocrystalline phases form composites with prominent increases in strength and plasticity. Use of MGs as the matrix for composites has gained significant attention for the development of advanced structural materials with unusual combinations of strength and toughness

[40, 46-48].

Extraordinary properties of MGs make them good candidate in different applications. In the next section some of the notable applications of MGs are introduced.

Literature Review

2.1.4 Applications of MGs

The unique properties make MGs a promising class of materials with various applications in vast areas, including electric transformers, fuel cells and sports elements [2, 13, 49, 50]. In this part, applications of MGs in biomedical nanotechnology and MEMs will be presented.

2.1.4.1 Biomedical applications

Metals and alloys have been widely used in biomedical applications [51]. Around

70% of the structural materials which are used in implants are metallic materials.

Hard-tissue prosthesis, bone screws, bone plates, artificial hip joints, knee joint and dental implants are among the biomedical applications of metals and alloys. Owing to unique characteristics, such as high strength, high elasticity and excellent wear and corrosion resistance, MGs have superiority over crystalline counterparts in some biomedical applications. Crystal defects in crystalline materials lead to the weakening of material, intergranular corrosion and stress-corrosion cracking in biological environments. Lack of these defects in MGs results in the superior performance in physiological environment [25]. Furthermore, in-vivo and in-vitro tests have confirmed the non-toxic interaction of MGs with cells and tissues [52].

Low Young’s Modulus compared with crystalline metals is another desirable property of MGs making them compatible with bones and reduces the risk of atrophy of bones. In addition, owing to their high flexibility and better compliance with blood vessels, MGs have been suggested for use as soft tissue stents. MGs are also

Literature Review

applicable for using as biodegradable implants. Recently Mg-based MG has been developed to be served as biodegradable implants [52].

2.1.4.2 Nanotechnology and MEMS

Similar to other types of polymeric and oxide glasses, MGs possess a SCLR where they exhibit excellent flexibility. Within this temperature window, MGs can be shaped like Newtonian viscous liquid under very small applied pressure [53]. This unique superplasticity and high stability make them ideal candidates in nanotechnology and MEMS [25].

In micro or nano machines as well as MEMS, three dimensional structure and surface pattern are widely required. These are conventionally produced by lithography and chemical etching, because nano-manufacturing of crystalline metals is expensive and lacks of excellent dimensional accuracy. Owing to superplasticity and lack of grain boundary, however, TPF of MGs provides a unique and economic method for fabrication of micro and nano products with excellent accuracy and surface finish in angstrom level [25]. Sub nanometer structural homogeneity makes patterned MGs suitable for optical applications as well. Sputtered MG films have been used for fabrication of micro cantilever in MEMS. Nanowires of MGs produced by TPF showed multi harmonic oscillations with potential applications in mechanical and magnetic fields. Some of the attractive properties of MGs as well as their potential applications have been presented in Tables 2.2 [5].

Literature Review

Table 2 .2 Properties and application field of MG [5]

Properties Application field High strength Machinery structural materials High hardness Cutting materials High fracture toughness Die materials High impact fracture energy Tool materials High fatigue strength Bonding materials High elastic energy Sporting goods materials High corrosion resistance Corrosion resistance materials High wear resistance Writing appliance materials High reflection ratio Optical precision materials High hydrogen storage Hydrogen storage materials Good soft magnetism Soft magnetic materials High frequency permeability High magnetostrictive materials Efficient electrode Electrode materials High viscous flowability Composite materials High acoustic attenuation Acoustic absorption materials Self- shaping property Penetrator High wear resistance and Medical devices materials manufacturability

It is interesting to note that how these alloys are fabricated for the mentioned applications. In the following section, major manufacturing techniques for MG production are studied.

2.2 Conventional manufacturing techniques of MGs

One of the key steps of commercialisation of any materials would be the manufacturing stage. Exploring an economic, environmentally friendly and reliable method for manufacturing of MG products has been an essential and highly applicable issue. Amorphous nature of these alloys creates some limitations in front of processing. The manufacturing processes should be designed and controlled in a

Literature Review

way that suppress crystallisation of MGs [54]. Crystallization occurs when cooling or heating patch intersects the crystallization curve in time-temperature- transformation (TTT) diagram [7]. The minimum cooling rate which must be employed in order to avoid crystallisation is schematically displayed as path 1 in Fig.

2.5. In this section two major conventional manufacturing processes along with their advantages and disadvantages will be discussed.

Figure 2.5 Cooling rate cycle for different manufacturing processes of amorphous alloys [7]

2.2.1 Casting

Casting has been the first and the most widespread manufacturing technique of amorphous alloys. In direct casting, the molten alloy is injected in to a mold and cools down rapidly to avoid crystallization (Path 1 Fig. 2.5). Die and suction casting

Literature Review

processes have been considered as net shape methods for MGs production [55].

However, because of lower porosity of the products, suction casting makes higher quality products compared with die casting. On the other hand, the advantage of die casting is its capability of high volume production of small to medium sized items.

Low melting temperatures of some MGs would be beneficial for die casting, as this issue would reduce the tool cost, wear, oxidation, energy consumption and casting time [7]. However, casting of MGs, similar to other alloys, would have some challenges, as well [55].

Shrinkage is a major issue in both die and suction casting processes. The shrinkage of a Zr-based MG is 0.5% approximately which is much smaller than crystalline counterparts. But, even the low shrinkage of MGs during casting causes a gap between the specimen and the molds. The gap is filled with a gas atmosphere (air) or vacuum with different heat conductivity that affects the cooling rate and amorphous structure of the products [7].

Moreover, because of the metastable nature of MGs and the requirement of high cooling rate for avoiding crystallization, the cooling and forming processes should be carried out simultaneously in casting. Casting is basically limited by two contractive requirements of devitrification and mold filling. Even by carefully balancing the casting parameters, limited simple shape products can be established by this method.

Furthermore, oxidation tendency of the melt at high temperature is the other factor impacting casting quality [6]. Accordingly, a careful balance of cooling rate for assuring mould filling and suppressing crystallization would be essential in casting of MGs. 25

Literature Review

The high viscosity of MGs would be another challenge in direct casting of MGs.

This issue is particularly challenging for complex shape and micro and nano-scale products. These challenges cause casting a low efficient method of manufacturing

MGs, especially at micro and nano scales [7].

2.2.2 Machining

Machining has been one of the major manufacturing processes of materials. Similar to other materials, machining has been considered as a manufacturing technique of

MGs. Improving the effectiveness of this process for MGs has been the topic of several research in the past two decades [8, 9, 56, 57]. In machining, tool life, surface finish and accuracy of the machined parts are among the most significant concerns [9].

Temperature plays a pivotal role in machining of MGs, as it directly affects the structure and flow behaviour of the alloys [56]. The thermal conductivity of MGs is very low (around 4 W/m K) by comparison with crystalline alloys. This issue make some challenges in front of their machining [8]. The low thermal conductivity leads to the temperature rise and higher oxidation or crystallization of the chips [8].

Strain rate is one further key factor in machining of BMGs. The workpiece-tool during machining is subjected to high strain rate deformation [8]. During machining of BMGs, even at very low rates, unique continuous chips with lamellar structure are formed [9]. The lamellar chips is suggested to be the result of activation of shear

Literature Review

bands in the structure of MGs owing to thermal instability [9]. Strain rate may also affect structure of BMGs. It was shown that even at very high strain rates, crystallization of BMGs during machining would be inevitable [8].

Except the mentioned concerns, other properties of BMGs including high hardness and wear resistance also creates significant barriers in front of efficient machining of these alloys [6].

2.3 Thermoplastic forming (TPF)

An alternative method of manufacturing MGs would be TPF. TPF provides a simple and economical technique for net-shaping of MGs [7]. By using TPF, precise and versatile geometries on wide length scales ranging from 10 nm to several centimetres have been produced which were previously unachievable with any conventional processing techniques utilized for crystalline alloys [58].

MGs due to their amorphous structure and existence of a SCLR between glass transition temperature and crystallisation temperature (Tx) can be thermoplastically formed [7, 58, 59]. This process is also known with other names such as superplastic forming, hot forming, hot pressing and viscous flow forming [7]. TPF of MGs was first introduced in 1978; however, until the discovery of amorphous alloys with desirable GFA, this technique was unexplored. Since then, TPF has been utilized for a wide range of processes such as micromolding, extrusion, blow molding, rolling, foam synthesis and fabricating ultra-smooth metal surfaces [6, 10, 54, 60]. In this technique, softening of MGs at SCLR allows these alloys to be shaped in processing

Literature Review

pressures which are comparable with plastics and not metals [7]. Meanwhile, forming process can be proceeding as long as the crystallization is inhibited. Higher

GFA reduces the risk of crystallisation in this method. GFA is typically evaluated based on the width of the SCLR. The larger the width of the SCLR, the higher the stability and GFA of the alloy [7].

Owing to the importance of TPF in this thesis, in the next three sections the procedure, applications and challenges in front of TPF will be discussed thoroughly.

2.3.1 Procedure

In TPF, the samples are heated to SCLR (path 2 Fig 2.5) where, MGs become soft and can be shaped under low pressure in a wide range of temperatures and time. At this region the viscosity of MGs would be in the range of 106-1010 Pa.s [6, 7].

Fig. 2.6 illustrates the schematic of TPF of BMGs. MG is placed on a die with desire feature and then both MG and the die are heated to the forming temperature and held for a period of time in that temperature to ensure temperature stabilization. At this stage the MG is a viscous liquid and ready for applying pressure. After applying pressure MG flow in to the die and the required feature can be established. Pressure, temperature and time are the primary factors playing key role in success of the process. If each of the mentioned parameters is not selected properly, the viscous resistance would not permit MG to flow in to the die and the process would be failed. Temperature plays the major role in TPF. Low temperature would increase the viscosity and reduces the formability of MG. On the other hand, high forming

Literature Review

temperature would raise the risk of crystallisation during forming. A typical forming temperature for TPF is chosen such that the magnitude of viscosity lies between 106-

108 Pa.s. As a general rule, in order to attain the maximum formability the highest temperature, pressure and time where the crystallization is avoided should be used in

TPF [6, 7].

Figure 2.6 Schematic of TPF of MGs

There is no requirement for fast cooling of MGs after TPF. TPF has been utilized in

MG product applicable in vast areas. Some of the applications are reviewed in the following paragraphs.

Microforming has attracted the researchers in a number of fields and their efforts lead to the fabrication of diverse parts ranging from MEMS components to micro fuel cells [50]. Microgear and nanowire applicable in MEMS and biomedical applications were fabricated using TPF [61, 62]. Surface patterning on MGs is another popular area which has applications in optical industries. Saotome [63, 64] pioneered imprinting of surface patterns on MGs. By utilizing TPF, MG thin-films, out-of-plane actuators and micro cantilevers were fabricated. The complex 3D 29

Literature Review

microstructures with high aspect ratios were thermoplastically formed from MG by

Bardt et al. [65]. They envisaged some possible applications including high-Q

(lightly damped) micro resonators for the telecom industry, high surface- area structures, microwave waveguides and connectors suitable for higher frequency operations, multi-degree-of-freedom flexure-based micromechanisms, microsurgical tools and devices, microscale motors and transmission components, microfluidic arrays, and free-form reflective micro optics [58].

TPF-based BMG micromolding was extended to the nanometer scale by taking advantage of wetting between the BMG and the die [6]. Through TPF-based nano- embossing, MG nanowires with very high aspect ratios (>200) were fabricated by

Schroers et al. [6, 53]. Fig. 2.7 illustrates the nanorods fabricated via TPF of a Pt- based MG on Si die.

Figure 2.7 Nanorod of Pt-based MG [62]

Literature Review

The nanowire architectures displayed superb durability combined with high electrocatalytic activity toward CO, methanol, and ethanol, exhibiting vast potential in fields such as energy conversion/storage and sensors [7]. Inoue noted the potential applications of the nanoscale-imprinted BMG surfaces in anti-reflection materials, cell culture media for bio-chips, electrode materials, hologram technology, and next- generation ultrahigh-density data storage material.

The TPF technique has also been employed to fabricate microlens arrays with potential applications in fuel cell interconnect plates. Owing to the unique mechanical properties of MGs, thermoplastically formed MG components have been used as a robust tool to replicate the micropatterns by microimprinting of the amorphous polymer poly(methyl methacrylate) (PMMA) [50, 58].

2.3.3 Challenges of TPF of MGs

Metastable structure of MGs makes some difficulties in front of TPF. MGs with a disordered metastable structure tend to become crystallized (ordered) under certain conditions such as high temperature, high strain rate and high pressure.

Crystallization significantly degrades properties of MGs, including anisotropy, excellent surface finish and formability. Thus, investigation of crystallization kinetics is of key importance in TPF processes in order to preserve the original MGs’ properties [58].

A schematic of time-temperature-transformation diagram (TTT diagram) has been illustrated in Fig. 2.5. Crystallization occurs when the processing path intersects the

Literature Review

crystallization curve. At lower temperatures (glassy zone), crystallisation time would be almost infinitive [6]. Within SCLR, however, crystallisation time is limited and varies based on the temperature [66]. This time varies significantly in a range of a few seconds until several hours, depending on the temperature [66]. A temperature dependent crystallization time and viscosity for a Zr-based BMG has been illustrated in Fig. 2.8 [7].

The viscosity of MGs is also basically controlled by crystallisation [66].

Crystallization increases the viscosity and reduces the formability of MGs during

TPF [66]. Change of amorphous structure of MGs can directly affect their properties.

However, thus far no research has been conducted in order to characterise the effect of TPF on the structure and mechanical property changes of MGs. Considering that the performance of a material in application is directly affected by their properties, improving or preserving the properties of MGs during TPF is essential.

Figure 2.8 Temperature dependent crystallisation time and viscosity for a Zr-based BMG [7] 32

Literature Review

Another major challenge in TPF of MGs is MG/dies adhesion. In TPF, dies play pivotal roles in production. Adhesion between workpiece materials (e.g., glasses [67] and polymers [68-70]) and dies causes the deterioration of the surface quality of products and dies, and has been a major challenge in TPF [68, 71]. Some methods have been utilized to separate MGs from dies, such as mechanical separation and dissolution of dies with chemicals [59, 62, 72, 73]. Nonetheless, they failed to fully resolve the problem, and often damaged either the MG components or the dies. Due to discontinuity of the physical properties and difficulties associated with experiments, structural and theoretical analyses of the interfaces have been always a challenging issue [74]. Mechanisms and work of adhesion are the major factors giving us valuable information in this area.

MG-die adhesion can be chemical, dispersive or/and diffusive [75], of which the intimate distance between the material pair is a critical factor. When an MG is heated to its supercooled liquid region, the material becomes a soft matter because of the dramatic decrease of viscosity. The low viscosity of the MG in this region leads to two key phenomena. First, the atom mobility increases rapidly, which enables them to easily rearrange themselves [76]. Secondly, the inter-diffusion and diffusion coefficients of the MG rise considerably [1]. These cause the MG to be able to wet the surface of another material in the supercooled liquid region [76]. Furthermore, because of the applying load during TPF, the distance between the atoms of the MG and a die surface is significantly reduced, which leads to the intimate contact between the material pair [77, 78]. All of these bring out challenges in selecting die materials for TPF of an MG. Wettability has been found to play an important role in

Literature Review

adhesion and it is strongly dependent on the surface energy and surface quality of a die material [79]. With increasing the wettability of a die surface, the contact area between an MG and the die extends. Materials with low surface energy, such as ceramics and some polymers [75], are hard to be wet and consequently are difficult to have a chemical reaction with other materials [80].

There exist several models investigating work of adhesion between materials [81,

82]. These models are basically based on Dupre equation and concerned with SFE of materials and surface interaction of the interfaces [82]. However, there is not any specific model determined the work of adhesion between MGs and other materials in

TPF. Consequently, the growth in the use of MGs in various applications, as well as lack of fundamental knowledge makes adhesion of MGs an essential issue and justifies the demand for further research.

In TPF of MGs dies should have specific requirements. The dies should be easily patterned and durable with reasonable strength and thermal resistance. A vast range of dies have been employed in TPF of MGs. In Table 2.3, the die materials with their specifications and properties have been exhibited.

Literature Review

Table 2 .3 Properties of manufacturing techniques of different dies in TPF [6]

Coeffficient Minimum Patterning of thermal Mechanical Material feature technique expansion behaviour size (1/˚C) Etchin Silicon >2 nm 3-5×10-6 Fragile electron beam Etchin Silicon dioxide >2 nm Strong electron beam Nickel Electroplating >1 µm 13×10-6 Strong Stainless steel Machining >50 µm 17×10-6 Strong Brass Machining >50 µm 20×10-6 Strong Carbon Pyrolysis FIB >1 µm 7×10-6 Strong Alumina Etching >10 nm 8×10-6 Strong SU8 Stamping >2 nm 52×10-6 Soft -6 Pt57.5Cu14.7Ni5.3P22.5 TPF >2 A 33×10 Strong

One of the most critical requirements of dies is the pattern resistance against collapse during forming process. The thermal expansion difference between dies and MGs is another factor that should be considered [35]. Large difference in thermal expansion leads to pattern distortion or stress build up in MGs and drastically affects the quality of products. Thermal expansion mismatch is also important in MG/die separation

[35].

Wetting between dies and MGs is a pivotal parameter in TPF, as well [83].

Especially in the features lower than 5µm, wetting takes an important part in TPF.

On one hand, high wettability of dies results in easy filling of the die by MGs and beneficial for TPF [83].

On the other hand, wetting promotes MGs/die chemical reaction which leads to severe adhesion. Adhesion between dies and MGs has been recognized as one of the pivotal issues in TPF. Regarding to its importance on the surface quality and cost of

Literature Review

both MGs and dies, MG/die adhesion consists one of the major sections of this thesis which will be discussed in details in Chapter 3.

Plastic deformation of MGs during TPF can help us better understand MGs behaviour and to address the challenges and ambiguities in this area. Accordingly in the next section the mechanism of plastic deformation of MG in glassy region and

SCLR are thoroughly investigated.

2.4 Mechanical behaviour of MG in TPF

The mechanical properties of materials play key roles in manufacturing of MGs.

Mechanical properties of MGs might be influenced by the processing technique [84].

Conventional metallic materials are crystalline in nature and their mechanical behavior is basically determined by the density of grain boundaries and dislocations and their ability to move [42, 84]. The presence of dislocations in crystals has also been considered to be the primary reason for their low strength (compared to the theoretical value) and ductile behavior [85]. Lack of the defects in their structures causes MGs to exhibit distinct mechanical behaviour. While their elastic modulus is in the same range of crystalline counterpart, their strength is much higher at room temperature [40, 44]. MGs do not exhibit any work hardening and their plastic deformation is influenced by both normal and shear stresses [40]. Owing to the absence of long range ordering, plastic deformation mechanism of MGs is distinct from crystalline materials [40, 44, 86]. In addition, deformation mechanism of MGs varies depending on the working conditions. While at room temperature, these alloys show no plasticity and they fail under tensile and compression loads, at SCLR they

Literature Review

convert to viscous flow with excellent formability. At glassy state, the deformation would be inhomogeneous and is localised in thin layers of atoms named shear bands.

However, at supercooled liquid state, the deformation is homogeneous and no shear band is formed during deformation [1, 40, 44]. Elastic, viscoelastic and viscoplastic are three major modes of deformation in MGs. At glassy region the first two contributions are activated. However, in SCLR viscoplastic contribution takes the pivotal part in the deformation [87].

Owing to the importance of mechanical properties in TPF and applications, in the following sections the plastic deformation of MGs both at glassy and supercooled state will be discussed. The conventional and new models in this area will be presented. Viscosity as the most notable parameter in TPF and its affecting parameters will be explored.

2.4.1 Inhomogeneous and homogeneous deformations of MGs

2.4.1.1 Inhomogeneous deformation

This mode of deformation is determined with creation and propagation of shear bands and leads to abrupt failure of materials [1, 88]. Shear banding or shear localization has been considered as the consequence of strain softening of MGs [1].

Strain softening during deformation is one of the characteristics of MGs [1, 44]. This means that an increase of strain makes the materials become softer and can be deformed by lower stresses [1, 7]. Shear bands and strain softening are considered to be the result of viscosity drop during forming [1, 88]. Different reasons have been

Literature Review

suggested for this phenomenon such as free volume creation due to flow dilatation, local evolution of structural order due to shear transformation zone (STZ), redistribution of internal stress as a result of STZ operation and local heating [1, 40].

However, two of them attracted more attentions in this area. The first hypothesis indicates that viscosity in shear bands decrease as a result of free volume production.

This results in the lower density and consequently lower resistance against deformation by comparison with the surrounding area [1].

In second hypothesis it is believed that local heating occurs in shear band which results in a viscosity drop by several orders of magnitude. This adiabatic heating could raise the temperature to higher than glass transition temperature and even melting point of MGs [1]. The temperature rise has been verified by analyzing stress strain curve of MGs [1].

Inhomogeneous deformation of MGs occurs at low temperatures and under high stresses [25, 89]. Shear bands approximately occurs in the plane with maximum resolved shear stress [90]. In tension, failure happens after the formation of the first shear band and because of this MGs show almost zero plasticity during tension. On the other hand, in constrained conditions such as uniaxial compression, indentation, bending and rolling, multiple shear bands operate leading to elastic, perfect plastic mode [43, 88]. Since shear bands carry large plastic strains, MGs show ductility in these conditions at room temperature. Nanoindentation is a perfect facility for characterizing the shear bands evolution in MGs [43].

Literature Review

In nanoindentation of MGs, serrated flow can been observed in load displacement curves, particularly at a lower strain rates [43]. Here also the activation of each shear band is accompanied by a drop in load displacement curve, called pop-in [43, 91]. At low loading rates, a single shear band is able to accommodate the deformation.

However, at high loading rates, multiple shear bands have to operate simultaneously to accommodate the deformation because of the short time. Because of this, the load displacement curve for former circumstances is serrated and for the latter would be smooth. As a result the pop-ins are more prominent in lower strain rate rather than higher ones [1, 43].

2.4.1.2 Homogeneous deformation

Homogeneous deformation of MGs typically occurs at high temperatures (>0.7 Tg) and SCLR. In these conditions, MGs can be considered as viscous flows with superplasticity behavior and have highly manufacturing importance. Although these alloys at room temperature inhomogeneously deformed with almost zero percent plasticity, at SCLR plasticity of MGs can be reached to as high as 1000% [86] .

A great deal of research has been done to analyze the effect of temperature, strain rate and stress on the homogeneous deformation of MGs [92-95]. It has been realized that at high temperatures and low stress level, these alloys behave very close to

Newtonian flows. In Newtonian regime, flow behavior is independent of strain rate.

However, with increasing stress and/or decreasing temperature, the flow becomes non-Newtonian and its behavior is strongly dependent on the stress and strain rate.

Literature Review

The major models describing flow and deformation behaviors of MGs in homogeneous mode which will be discussed in the following sections [1].

2.4.2 Theoretical model of plastic deformation

As the chemical bonds in MGs are of primarily metallic character, plastic strain can be easily accommodated at the atomic level [96]. However, due to its cluster- assembled structure, the deformation mechanism of an MG is absolutely dissimilar to that of conventional crystalline alloys. In this section, the progressive understandings on the particular deformation mechanism of MGs are discussed. Free volume and STZ as the major models in plastic deformation of MGs is discussed thoroughly.

2.4.2.1 Free volume model

Unlike the aggregated dislocations in crystalline metals, the plasticity events in MGs do not leave apparent structural imprint. In the 1960s, Turnbull and co-workers [97,

98] proposed a “free volume model” for understanding glass transition. Spaepen [99] further applied this model to the case of glass deformation. The concept of this model came from the observation of the low density of MGs compared with their crystalline counterparts. As such, around atoms additional space or free volume should exist. The underlying deformation of MGs is thus expected to be through an atomic jump from a high density zone to a new site with lower density [3]. A schematic of atomic jump based on free volume model is shown in Fig. 2.9.

Literature Review

Figure 2.9 Atomic jump in free volume model [44]

The free volumes in MGs are described as density fluctuations with the volume

* greater than a critical value (υ ). In this way, the free volume concentration (cf) is evaluated by [100-102]:

= exp = exp ( ) (2-1) 𝛾𝛾0𝜐𝜐∗ 1 𝑐𝑐𝑓𝑓 � 𝜐𝜐𝑓𝑓 � 𝑥𝑥 where is a geometrical overlap factor ranging from 0.5 to 1 and is the average

0 𝑓𝑓 free volume𝛾𝛾 per atom. The quantity x is the reduced free volume𝜐𝜐 . Free volume concentration is not a constant value and varies by two distinct phenomena, structural relaxation and plastic deformation.

Structural relaxation is a consequence of structural fluctuations in MGs and is a function of temperature and plastic deformation [87]. Structural relaxation is accompanied by changes in physical and mechanical properties, such as density and hardness [87]. Dynamic mechanical analysis is an efficient method for characterising structural relaxation in MGs. There are two main types of relaxations in MGs, the primary relaxation (α) and secondary relaxation (β). Primary relaxation is accompanied by the rearrangement of a large group of atoms and occurs at high temperatures (supercooled) or low frequencies. On the other hand, a few atoms

Literature Review

contribute in secondary relaxation mode and rearrange locally. Secondary relaxation in contrast with primary one takes place at low temperatures (glassy) or high frequencies [87, 103, 104].

Structural relaxation annihilate free volumes as a result of inducing ordering via diffusion [87, 100]. Based on free volume model, raising temperature increases the diffusion rate and facilitates structural relaxation [105, 106]. The rate of free volume annihilation is calculated via:

= (2-2) 𝑓𝑓 𝑑𝑑𝑐𝑐 2 𝑑𝑑𝑑𝑑 −𝑘𝑘𝑟𝑟𝑐𝑐𝑓𝑓 where kr is a constant. According to Eq. (2-2), free volume concentration of MGs decreases with time owing to structural relaxation.

On the other hand free volume can be created during plastic deformation under high stress and strain rates [101]. Free volume creation can be modeled based on the density fluctuations with the volumes smaller than the critical size υ*. Under stress, the neigbouring atoms can be squeezed and creates free volume [102]. The free volume creation can be expressed as:

= (2-3) 𝑑𝑑𝑐𝑐𝑓𝑓 𝑑𝑑𝑑𝑑 𝑐𝑐 𝑓𝑓 where is a proportionality constant. By combining𝑘𝑘 𝑐𝑐 equations 2-2 and 2-3 the total

𝑐𝑐 free volume𝑘𝑘 annihilation and creation during TPF forming of MGs can be calculated by:

= + (2-4) 𝑓𝑓 𝑑𝑑𝑐𝑐 2 𝑑𝑑𝑑𝑑 −𝑘𝑘𝑟𝑟𝑐𝑐𝑓𝑓 𝑘𝑘𝑐𝑐𝑐𝑐𝑓𝑓

Literature Review

According to free volume model, structural relaxation and deformation induced free volume play the pivotal role in the deformation of MGs. At high temperatures and low stress level, free volume creation as a result of deformation is negligible and structural relaxation would be the key factor. On the other hand, at high stresses, the free volume creation is prominent and strain softening occurs in MGs, accordingly.

2.4.2.2 STZ model

In contrast with free volume, in STZs model, deformation is based on the collective rearrangement of the atomic cluster rather than a simple jump. On the basis of a bubble-raft model, Argon et al. [107] proposed a STZ model, as schematically shown in Fig. 2.10. In this theory, the deformation mechanism of MGs in the SCLR is based on the local atomic rearrangement of the fundamental flow unit called STZ

[44]. STZs are local cluster of atoms that under shear stress undergo from one low energy configuration to a second such configuration in PEL (See Fig. 2.2) [44]. It should be noted that STZs are not structural defects in MGs and have a transience nature [44]. It is believed that STZ contain 100~200 atoms.

Literature Review

Figure 2.10 Movement of a STZ during applying shear stress [108]

After discussing the plastic deformation mechanism of MGs in SCLR, in the next section viscosity as a primary parameter in TPF is studied.

2.4.3 Viscosity

Viscosity is the one of the most important physical property of MGs in SCLR which determines their GFA and flow behaviours [109]. Viscosity is defined as the resistance of a liquid to shear stress. Liquids, depending on the viscosity variation with temperature, have been classified to two categories, fragile and strong. In strong liquids, the viscosity is high near melting point and it increases with decreasing temperature with an Arrhenius type equation. In fragile liquids, viscosity is low near melting point and increases slowly with decreasing temperature. Fragile liquids do not follow the Arrhenius equation and can be better explained by Volger-Fulcher-

Tammann (VFT) equation [1]. MGs are intermediate fragile liquids and VFT equation better evaluates their viscosities at each temperature [1, 110].

Literature Review

According to its relation to shear flow, it is plausible to determine the viscosity by measuring the flow stress at different strain rates by using the following equation:

= (2-5) 𝜎𝜎𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝜂𝜂 𝛾𝛾̇ where is viscosity, is flow stress and would be shear strain rate. Different

𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 experimental𝜂𝜂 methods𝜎𝜎 are also available for𝛾𝛾 ̇determination of viscosity including decay of resonance oscillation [109], parallel plate rheomentry [111], three point beam bending [112] and length changes of glassy ribbons [113].

In MGs the viscosity is very high until SCLR. Near the glass transition, the viscosity of an MG decreases sharply, which could be up to 14 orders in a relatively narrow range of temperature. At this region, the viscosity drops slowly and reaches to a minimum value. The minimum viscosity indicates the crystallisation temperature and after Tx viscosity increases until it reaches to a steady state value.

The flow behaviour of the MGs plays an important role in the TPF of MGs. It was confirmed that non-Newtonian flow promotes the crystallization at a specific temperature and limits TPF ability [114]. On the other hand, Newtonian flow in TPF improves the formability of MGs. A vast number of studies have been done in order to investigate the effect of strain rate and temperature on viscosity [86, 115-118] and will be discussed in the following sections.

Literature Review

2.4.3.1 Temperature effect

The variation of viscosity with temperature is of great importance in understanding

MG behaviours and technological applications. As discussed earlier, temperature dependent viscosity of MGs can be calculated via different models such as Arrhenius and VFT [1]. However, it believes that VFT equation predicts the viscosity of MGs more accurately:

= ( ) ∗ (2-6) 𝐷𝐷𝑇𝑇 ∗ 𝜂𝜂 𝜂𝜂0𝑒𝑒𝑒𝑒𝑒𝑒 𝑇𝑇−𝑇𝑇 -5 where η is viscosity at temperature T and η0 is a constant equal to 4×10 and D and

T* are fragility parameter and and VFT temperature, respectively [110].

Viscous property of liquids is typically determined by fragility parameter (m). A high fragility parameter value corresponds to fragile fluids and a small value corresponds to strong liquids, nearly Arrhenius type behaviour. Fragility values for

MGs are typically around 30-60 indicating intermediate fragile character for MGs

[110, 119]. Fig. 2.11 displays a typical viscosity-temperature change for a variety of strong to fragile liquids.

Literature Review

Figure 2.11 Viscosity variation with temperature for a number of amorphous materials [1]

Viscosity dependent temperature directly affects thermoplastic formability of MGs.

Under similar condition of environment, time and load, the thermoplastic formability of MGs increase with temperature as long as crystallisation can be avoided [66]. A number of researches have been done on the thermoplastic formability of MGs in order to relate this factor to one of the known characteristics of MGs. Once, it was thought ΔT=Tx-Tg would be the best criterion in defining the thermoplastic formability of MGs [64]. However, recently a new factor S= has been ΔT 𝑇𝑇𝑙𝑙−𝑇𝑇𝑔𝑔 introduced with higher efficiency in determining the formability of MGs where Tl is liquidus of the MG [120].

2.4.3.2 Strain rate effect

Strain rate can be considered as one of the key parameters impacting on viscosity, relaxation and TPF of MGs [59]. It is widely accepted that increasing strain rate

Literature Review

facilitate the transition from Newtonian to non-Newtonian state in homogeneous deformation [59]. The change of flow behaviour has been demonstrated to have a strong impact on the deformation mechanism of MGs. In Newtonian state, the viscosity is independent of strain rate and its behaviour is mainly dependent on temperature. However, increasing strain rate in non-Newtonian state reduces the viscosity and makes MGs softer as a result of free volume creation.

It is believed that strain rate and free volume concentration is directly proportional.

Increasing strain rate raises the free volume concentration. Besides, with regard to viscosity, flow stress and strain rate relationship (η=σ/ ) one can understand that viscosity in inversely proportional to free volume concentration.𝜀𝜀̇

2.4.3.3 Apparent viscosity in TPF

Apparent viscosity is widely used in TPF of MGs. This term is used to quantify the viscosity of MGs at TPF. This term in fact includes the internal and external factors that may affect viscosity, such as friction, structure, pressure and die size. Apparent viscosity would have high scientific and technological importance and because of this, the mathematical model that evaluates apparent viscosity is presented as well as reviewing some literature in this area.

A number of research has been done on micro and nano moulding of MGs under different conditions. Shao et al. fabricated Pt-based MG nano rods at different sizes and found that the apparent viscosity is significantly affected by the size of the dies especially at lower than 100nm. They found that at lower than 100 nm sizes, the

Literature Review

apparent viscosity of MGs increases significantly. The exact reason behind this behaviour has not been mentioned in this paper, but the change of the flow behaviour in nanometer scales has been considered as the most probable reason.

Wu et al. [121] investigated the viscosity dependence of a Zr based MG in SCLR by designing an extrusion process with different inner and outer die ratios. They found that by changing the ratio of inner to outer diameter of the cups, the apparent viscosity of the material changed. They contributed the apparent viscosity change to the friction change of the die.

On the other hand, some papers repeated that apparent viscosity is independent of the dies sizes and remains constant at different dies sizes [6, 73, 122]. Other factors such as oxidation effect and capillary forces on apparent viscosity has been explored in

[123, 124]. They showed that oxide layer can act as a barrier in front of TPF of MG.

Thus apparent viscosity is key factor in TPF of MGs. This factor would be considered in chapters 5 and 6 for two different die features.

2.5 Summary

MGs are alloys with amorphous structure. Owing to lack of defects and any long range ordering in their structures, MGs possess exceptional properties such as high elastic limit, hardness, wear and corrosion which make them good candidates in many applications. Because of non-equilibrium structure and brittleness, manufacturing of these alloys via conventional techniques such as casting and machining has been challenging. However, due to superplasticity of these alloys

Literature Review

within SCLR, TPF has been recognized as the most efficient manufacturing technique of MGs. Meanwhile, TPF has been faced with serious challenges that should be resolved in order to make it a costly effective and reliable method.

Adhesion of MGs and dies has been among the major problems in this area. Some methods such as dissolving dies in a chemical or mechanical separation have been utilized which deteriorate the die and/or MG surfaces and significantly reduced the applicability of TPF. In addition, mechanism of adhesion in TPF is still ambiguous and there is no model to quantitatively estimate the work of adhesion between die and MGs.

Temperature as the main factor in TPF of MGs has been studied in different research. It is recognized that increasing temperature as long as crystallisation is inhibited can improve the formability of MGs. However, a systematic study on the effect of temperature on viscosity evolution of MG in TPF is still lacking. Knowing the viscosity evolution can reveal the materials behaviour during TPF.

The effect of TPF on the structure of MGs is another area that requires further investigations. It is known that MGs might be crystallised at high temperature.

Though the structure of MGs by using XRD and TEM has been studied, the effect of

TPF on the structure of MGs is lacking.

Another important area that still remains unclear is the effect of TPF on the mechanical properties of MGs. There are several of research discussing the effect of stress, strain rate or temperature on the stress-strain or viscosity during TPF. 50

Literature Review

However, a systematic study on the effect of TPF parameters, such as temperature and/or die on mechanical properties is still lacking. There exists an ambiguity that whether or not TPF influences the mechanical property of MG. This is extremely important as it directly affects the performance of MGs in service. Accordingly, examination of mechanical properties after TPF for revealing the property changes during processing is essential.

In conclusion, in this thesis MG/die adhesion in TPF and the effect of TPF parameters on the apparent viscosity, structure and mechanical properties of MGs are thoroughly examined in the next chapters.

Die Selection

Chapter 3 Die Selection

Die Selection

As mentioned in Chapter 2, adhesion between MG and die is a major issue in TPF of

MGs and increases the net price of the products significantly owing to deterioration of die or MGs. Accordingly, MG/die adhesion in TPF is thoroughly investigated in this Chapter. Several materials are selected and utilized as die. After TPF, microscopic and elemental analyses are conducted to identify the mechanism of adhesion in this process. The primary material parameters influencing the adhesion status of the dies are recognized. A new model is introduced for estimation of SFE of

MGs. A novel model is developed for evaluation of the work of adhesion between

MG and die in TPF and verified by the experimental results.

3.1 Materials selection

Two MGs, La-based and Zr-based, were chosen in this study for the investigation of their adhesion with a number of commonly used die materials. The La-based MG,

La60.5Al16.3 (Cu, Ni)23.2, was prepared by arc melting of high purity elements (>

99.5%) under the Ti-gettered argon atmosphere, followed by suck casting into copper mould. To produce a homogeneous product, the ingot was remelted several times. The Zr-based MG, Zr58.5 Cu15.6 Ni12.8 Al10.3 Nb2.8 (LM106a), was purchased from LIQUID METAL TECHNOLOGY. Differential scanning calorimetry (DSC) test was done for measuring the glass transition and crystallization temperatures of the La-based MG by Perkin Elmer DSC 7 at a heating rate of 20˚C/min (Fig. 3.1). X-

Ray diffraction (XRD) analysis was carried out for verifying the amorphous nature of La-based MG by using Cu kα source (Fig.3.2). For the LM106a, the data supplied by LIQUID METAL TECHNOLOGY was used as a reference. Based on the experiments and the data available, the glass transition and crystallisation of

Die Selection

La60.5Al16.3 (Cu, Ni) 23.2 were determined to be 145 ˚C and 205 ˚C, respectively. For the Zr-based MG, these were 390˚C and 497˚C, respectively. The XRD analysis of

LM106a is shown in Fig. 3.3 which verifies the amorphous structure of the sample.

1.5 1.3

1.1 0.9 0.7 0.5 Tx 0.3 Tg 0.1 Heat Heat flow (W/g) -0.1 -0.3 -0.5 0 50 100 150 200 250 300 350 Temperature (˚C)

Figure 3.1 DSC result of La60.5Al16.3 (Cu, Ni)23.2

3000

2500

2000

1500

1000 Intensity (Counts) 500

0 30 40 50 60 70 2θ (deg)

Figure 3.2 XRD results of La60.5Al16.3 (Cu, Ni)23.2

Die Selection

2500

2000

1500

1000

Intensity (Counts) 500

0 20 30 40 50 60 70 2θ (deg)

Figure 3.3 XRD result of Zr58.5 Cu15.6 Ni12.8 Al10.3 Nb2.8

Seven die materials selected in this study were electroless Ni-P, polytetrafluoroethylene (PTFE), Si, sapphire, SiC, Si3N4 and WC-Co. It was noted that PTFE and electroless Ni-P were only applicable to La-based MG due to the low temperature required for their thermoforming processes [125].

The electroless Ni-P die used in this study was an amorphous coating with a thickness of 100 µm on copper, fabricated by electroless method. Due to its advanced tribological properties and machinability, electroless Ni-P dies have been widely used in the manufacture of precision products in nano and micro scales [126].

Si dies are among the most popular dies in MEMS and electronic circuit industries, owing to the simplicity of fabricating micro and nano scale features, and desirable mechanical and electrical properties [72, 73]. In this study Si wafer with around 600

µm in thickness was used as a die material. PTFE is an anti-sticking polymer material, which has a very low surface energy and wettability. In this study, a layer of PTFE with around 100 µm in thickness was coated on copper substrate using 55

Die Selection

PTFE tape. Sapphire, SiC, Si3N4, alumina and tungsten carbide (WC) have outstanding tribological properties as a result of their high hardness, high wear resistance and low friction coefficients and thus have a wide range of applications in various industries [127-131]. Thus these materials were also selected for investigation in this thesis.

3.2 Experimental set up

The TPF processes were conducted on a Toshiba precision glass moulding machine

(GMP-211) in the Nano and Precision Engineering Lab at the UNSW Australia (Fig.

3.4). All of the tests were carried out at a constant moulding temperature (160 ˚C for

La-based and 460 ˚C for Zr-based MGs) under the moulding load of 200 N and loading time of 180 s. The temperature, load and loading time were selected according to two phenomena: crystallisation and Newtonian flow. The minimum applicable load was selected in all experiments to ensure the MGs behave Newtonian during the experiment. The temperature and loading time were also chosen to have sufficient formability and time during tests. At higher temperatures, the loading time becomes small and at lower temperatures, the MGs did not have sufficient formability.

To remove the effect of shape on adhesion, flat dies were used at all experiments.

MGs had a circular shape with 2.5 mm diameters and around 1 mm thickness.

Before the TPF tests, the surfaces of the MG samples were ground and polished to a surface roughness of Ra ~ 100 nm. After the TPF processes, the adhesion status of the different MG-die material pairs were investigated by means of optical microscope, scanning electron microscope (SEM model Hitachi 3400I), high

Die Selection

resolution SEM (FEI NOVA Nano SEM 450) and energy dispersive spectroscopy

(EDS).

Figure 3.4 Toshiba precision glass moulding machine (GMP-211)

3.3 Adhesion status of different dies and MGs

In the following sections the adhesion status of MGs with a number of dies is investigated.

3.3.1 Die made of electroless Ni-P

Electroless Ni-P is an amorphous material. Due to its excellent wear resistance, it has been used as a die material for the TPF of polymers and glass [132]. Nevertheless, when it is used for the TPF of La-based MG, some materials remained on the die

Die Selection

surface after demoulding (Fig. 3.5a). The EDS image in Fig. 3.5b confirmed that the residual materials were the workpiece material, La-based MG.

Figure 3.5 (a) SEM image of the electroless Ni-P die surface after TPF, and (b) the EDS result of the residual materials on the die surface after demoulding

3.3.2 Die made of the Si wafer

Both the La-based and Zr-based MGs could easily wet the surface of Si and bond to it. To study the bonding interface, the bonded Si and La-based MGs were cut transversely and the interfaces were observed by means of optical microscopy (Fig.

3.6a) and high resolution SEM (Fig. 3.6b). As shown in Fig. 3.6b, a distinct layer with a thickness of 240 to 370 nm formed between the Si die and La-based MG. In order to obtain the approximate chemical composition and the elements in this layer,

EDS point and mapping analyses were conducted (Fig. 3.7 a and b). The figure clearly shows that the interface contains a mixture of Si and some elements of the

La-based MG. Fig. 3.7a also indicates the diffusion between Si and MG.

Die Selection

Figure 3.6 (a) Optical microscope image of the La-based MG and Si die, and (b) high resolution SEM image of the interface layer

Figure 3.7 (a) EDS mapping for the La-based MG, interface and Si after TPF, (b) EDS point analysis of the interface layer between the La-based MG and Si

3.3.3 The PTFE die

Figure 3.8 shows the surface of the La-based MG after TPF when using the PTFE die. As can be seen, no bonding occurred between the MG and PTFE. However, due to low melting temperature of PTFE (320 ˚C), this material is only applicable to the

MGs of low glass transition temperature such as La-based or Mg-based.

Die Selection

Figure 3.8 SEM image of the La-based MG after TPF on a PTFE die

3.3.4 SiC

Figures 3.9 a and b show the die surface after TPF of a La-based and Zr-based MGs on SiC, respectively. SiC have medium range of adhesion with MGs.

Figure 3.9 Optical microscope images of the SiC die surface after TPF with (a) the La-based MG, and (b) the Zr-based MG

Die Selection

3.3.5 Si3N4

Figure 3.10 shows the surface of Si3N4 ceramic after TPF. It is evident that adhesion happened between La-based MG and die. The extent of adhesion seems to be higher than SiC but lower than Si and electroless Ni-P.

Figure 3.10 Optical microscope image of the Si3N4 die after TPF with the La-based MG

3.3.6 Alumina

The surface of alumina after TPF by La-based MG has been illustrated in Fig. 3.11.

The extent of adhesion for alumina compared with other dies can be categorized as slight to medium range.

Die Selection

Figure 3.11 Optical microscope image of the alumina die after TPF with the La- based MG

3.3.7 Sapphire

Figures 3.12a and 3.12b show the surfaces of sapphire die after TPF process for La- based and Zr-based MGs, respectively. The adhesion resistance of the sapphire die is the best except PTFE. Slight bonding happened in only very tiny spots with the

MGs.

Die Selection

Figure 3.12 Optical microscope image of the sapphire die after TPF with (a) the La- based MG, and (b) the Zr-based MG

3.3.8 WC-Co

WC-Co die which is known as cemented carbide is a composite where WC particle distribute in Co as a binder and is a materials which is widely used in metal forming due to its attractive tribological properties, high shock resistance and high melting temperature [129, 133]. The structure of WC-Co is a hexagonal structure which is stable until 2000 ˚C [133]. It is apparent in Fig. 3.13 that slight adhesion took place between Zr-based MG and WC-Co after TPF. Thus, WC-Co showed desirable resistance against adhesion with MGs.

Die Selection

Figure 3.13 Optical microscope image of the WC-Co die after TPF with Zr-based MG

3.4 Adhesion mechanism of dies and MG

Table 3.1 summarises the SFE, forming temperature and adhesion statues of the employed materials with the MGs after TPF. It can be seen that the selected materials have a wide range of SFE and the extent of adhesion varied significantly among them. WC-Co is a metal matrix composite where Co plays as the binder. The

SFE data for WC-Co could not be found in the literature therefore the SFE value has not been reported in Table 3.1. PTFE, WC-Co and sapphire showed the best performance against adhesion and electroless Ni-P and Si showed the highest bondability. The status of bondability of SiC, alumina and Si3N4 were between the former and latter groups. Fig. 3-14 compares the bonding dissociation energy (BDE) of the primary constituent bonds in each material [134], where the BDE is the energy required for breaking a bond. W-W and C-C were the primary bonds in WC-Co. C-C

Die Selection

and C-F are the primary constituent bonds in PTFE; and Al-O, Si-C and Si-Si are the major bonds in sapphire, SiC and Si, respectively. In electroless Ni-P, there are three types of bonds, including Ni-Ni, Ni-P and P-P. Considering the high percentage of

Ni (90%) in the material structure of electroless Ni-P, the BDE of Ni-Ni was thus used as the average BDE of the electroless Ni-P in this study. The adhesion results had a very good agreement with the BDE of the primary constituent bonds of the materials, showing that the higher the BDE, the less MG-die adhesion took place.

For WC-Co and PTFE which had the highest BDE of the constituent bonds, no adhesion occurred. For the others, bonding area decreased with increasing the BDE.

The detailed adhesion behaviour and mechanism of each pair will be discussed in the following sections.

Die Selection

Table 3.1 Summary of adhesion status and the SFE of MGs and die materials

Surface energy of Metallic Die Temperature Microscopic the die glasses Materials (˚C) adhesion level materials (mJ/m2) La- Electroless 1093[135, based 160 Medium Ni-P 136] MG La- based Si 1250[135] 160 Severe MG La- PTFE on based 18.5 [137] 160 No adhesion copper MG La- based Si3N4 ~900 160 Medium to severe MG La- based Alumina 905 160 Medium MG La- based Sapphire 638 [138] 160 Slight MG La- 905[135, based SiC 160 Medium 139] MG Zr- based Si 1250[135] 460 Severe MG Zr- based Alumina 905 460 Slight to medium MG Zr- based Sapphire 638 [138] 460 Slight MG Zr- 905[135, based SiC 460 Slight to medium 139] MG Zr- based WC-Co - 460 Slight MG

Die Selection

700

600

500

400

300

200

100

Bonding dissociationenergy (KJ/mol) 0 W-W C-C C-F Al-O Si-N Si-C Si-Si Ni-Ni Primary bond of the die

Figure 3.14 BDE of the constituent bonds of the die materials selected for investigation

The weak chemical bonds and relatively high SFE along with the high tendency of chemical reaction between the elements of the La-based MG and Ni-P result in the adhesion. Particularly, the affinity of Ni, Cu and La atoms is high [140, 141] and these elements are able to react chemically and promote the adhesion [140, 141].

Considering that La, Cu and Ni are the main elements in the structure of the La- based MG, they possess high tendency for mutual solubility with Ni atoms of the electroless Ni-P [140] and thus lead to the observed adhesion. Moreover, the diffusion coefficients of elements in the MGs in the SCLR are much larger than those in glassy state and thus these atoms are able to diffuse into other materials easily in this region. It has been proved that the diffusion coefficient of Ni atoms in a

Zr-based MG is several orders of magnitude higher than the glassy state [1]. As a

Die Selection

result, diffusion bonding and metallurgical reaction should be responsible for the adhesion between La-based MG and Ni-P.

Si can form stable products with either La or Zr atoms, if certain temperature and pressure conditions are satisfied [142, 143]. Si has a relatively high SFE and thus its surface can be effectively wetted by the MGs in their SCLR [83]. This causes the

MGs to find the intimate distance with Si atoms, leading to chemical reaction and forming a new compound at the interface (Fig 3.6b). The creation of the new compound is evidenced by the EDS analyses in Figs. 3.7 a and b showing a unique mixture of Si (die) and La-based MG elements at the interface layer. Moreover, during TPF, their atoms can diffuse into each other, as evidenced in Fig. 3.7a, owing to the high diffusion coefficients of MGs and Si. Therefore, it could be concluded that chemical adhesion and diffusion bonding are responsible for the MG-Si adhesion.

PTFE is a fluorocarbon and its molecular formula is (C2F4)n produced from the polymerization of the tetrafluoroethylene. C-F and C-C bonds are the major bonds in

PTFE and both of them are among the strongest single bonds in organic chemistry.

Due to this structure, PTFE has a very low SFE [144, 145] and is one of the most stable organic compounds [146]. Therefore, wetting of PTFE by other materials is difficult, making it an ideal anti-wetting layer in TPF process to avoiding bonding/adhesion [147]. Owing to the lack of intimate distance with MGs, a PTFE die

surface eliminates the probability of chemical adhesion and diffusion bonding [75].

Furthermore, it is evident from Fig 3.15a that BDE of C-C is much higher than either 68

Die Selection

Zr-C or La-C bonds; hence it is highly unlikely that these atoms are able to replace C in the structure of PTFE. A further possibility is the substitution of Zr or La atoms instead of F in the C-F bonds. However, this can be removed, as the C-F bonding energy (Fig. 3.14) is higher than any of the potential replacement bonds. As a result,

MGs would remain their cohesion, rather than adhesion with PTFE.

For the sapphire, alumina, Si3N4, WC and SiC dies, the mechanisms of adhesion with the MGs are almost the same. All the materials contain covalent and ionic bonds which are very strong bonds. The SFEs of sapphire is lower than the others.

However, the other dies including alumina, Si3N4 and SiC possess almost the same

SFE (Table 3.1). However, the BDE of the constituent bonds of sapphire is higher than that of SiC (Fig. 3.14). Hence, the bonds in sapphire are more stable and the likelihood of chemical reaction between sapphire and the MGs is lower. Figs. 3.15b and 3.15c show the BDEs of the primary bonds in SiC and sapphire, respectively.

For comparison, the BDEs of the constituent bonds of oxides and carbides of La and

Zr were also included in the figures. Al-O is the primary bond in sapphire and alumina. From the thermodynamics point of view [148], La-O and Zr-O form more stable bonds compared with Al-O and therefore the chemical adhesion between sapphire and MGs would be possible under certain conditions. However, sapphire because of lower SFE compared with alumina, showed less adhesion. The similar chemical reactions are even more likely to occur with SiC because of its weaker bonds. Thus the MG-SiC adhesion is higher than that of the MG-sapphire. In Si3N4 also Si-N bonds have lower energy than La-N and Zr-N, which make their reaction thermodynamically possible (Fig. 3.15 d). Cement tungsten carbide has the highest

BDE among the employed dies. Although finding average BDE value for WC-Co is 69

Die Selection

difficult due to its complex composite structure but W-W and C-C which are the major bonds in this material have the highest value among the dies. WC-Co therefore has a very stable structure and breaking its atoms and replacing by another element would be very difficult. Lack of enough data about its SFE makes it hard to predict the wetting of MG on its surface, but experimental results by using optical microscope confirmed its good resistance against adhesion.

Figure 3.15 BDE of (a) C-C, Zr-C and La-C, and (b) Si-C, La-C and Zr-C (c) Al-O, Zr-O and La-O (d) Si-N, La-N and Zr-N

Die Selection

3.5 SFE of MGs

SFE is an important factor in TPF of MGs at micro and nano scales [123]. No model has been developed for calculation of SFE of MGs. Here a model is presented based on macroscopic atom model to calculate SFE.

3.5.1 A model for calculation of SFE of MG

Macroscopic atom model has been widely used for calculation of SFE and work of adhesion between metallic systems [81]. According to Refs. [81, 135, 136], the surface energy of a compound AxBy can be expressed as:

C S ∆H surf C S ∆H surf γ = A A + B B Ax By A 2 / 3 B 2 / 3 (3-1) fvacuumc0VA fvacuumc0VB

S S where C A and CB are the surface area fractions of the component A and B, respectively. Surface area fraction among a pair of elements is calculated based on

surf their atomic radius and electronegativity and details can be found in [81]. ∆H A and

surf ∆H B are the surface enthalpy of A and B, respectively, and their values have been

A B measured for majority of the elements of the periodic table [135]; fvacuum and fvacuum are the degrees to which the surface atoms A and B are surrounded by vacuum [135], and generally their average value 0.31 were used in the calculation [135]; VA and VB are the molar volumes of atoms A and B in their pure state, respectively, and c0 is a

8 constant related to Avogadro number and its value is around 4.5 × 10 [81].

Considering the amorphous structure of MGs, the SFE of the amorphous metals can be roughly considered as almost 50% of the values of the crystalline alloys with the

Die Selection

same composition [136]. Eq. (3-1) can thus be extended to the calculations of γMG.

Take Zr58.5 Cu15.6 Ni12.8 Al10.3 Nb2.8 for example, its SFE can be expressed as

. . . . .

𝑍𝑍𝑍𝑍58 5𝐶𝐶𝐶𝐶15 6𝑁𝑁𝑁𝑁12 8𝐴𝐴𝐴𝐴10 3𝑁𝑁𝑁𝑁2 8 𝛾𝛾 1 = { + + 2 𝑠𝑠 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑠𝑠 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑠𝑠 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝐶𝐶𝑍𝑍𝑍𝑍∆𝐻𝐻𝑍𝑍𝑍𝑍 𝐶𝐶𝐶𝐶𝐶𝐶∆𝐻𝐻𝐶𝐶𝐶𝐶 𝐶𝐶𝑁𝑁𝑁𝑁∆𝐻𝐻𝑁𝑁𝑁𝑁 2 2 2 𝑍𝑍𝑍𝑍 �3 𝐶𝐶𝐶𝐶 �3 𝑁𝑁𝑁𝑁 �3 𝑓𝑓𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑐𝑐0𝑉𝑉𝑍𝑍𝑍𝑍 𝑓𝑓𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑐𝑐0𝑉𝑉𝐶𝐶𝐶𝐶 𝑓𝑓𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑐𝑐0𝑉𝑉𝑁𝑁𝑁𝑁 + + } 𝑠𝑠 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑠𝑠 𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝐶𝐶𝐴𝐴𝐴𝐴∆𝐻𝐻𝐴𝐴𝐴𝐴 𝐶𝐶𝑁𝑁𝑁𝑁∆𝐻𝐻𝑁𝑁𝑁𝑁 2 2 𝐴𝐴𝐴𝐴 �3 𝑁𝑁𝑁𝑁 �3 𝑓𝑓𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑐𝑐0𝑉𝑉𝐴𝐴𝐴𝐴 𝑓𝑓𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑐𝑐0𝑉𝑉𝑁𝑁𝑁𝑁 (3-2)

The calculated SEFs for Zr-based MG and La-based MG are shown in Table 3.2, which agrees very well with the reported values [62, 72, 83, 136]. It is clear that the surface energy of La-based MG is lower than Zr-based MG. As Cu, Al and Ni are common elements with almost the same atomic percentage in both MGs ( 38.7% in

Zr-based and 39.5% in La-based MG) and the percentage of Nb in LM106a is very low (2.8 atomic%), La and Zr atoms should play the key role in SFE of La-based and Zr-based MGs as well as their bondability, respectively. Considering that the

SFE of pure La (1020 mJ/mm2) is almost half of the SFE of pure Zr (2000 mJ/mm2)

[135], It is therefore quite reasonable that with regard to the same percentage of these elements in MGs, the SFE of Zr-based MG becomes much higher than La- based MG.

The calculated SFE of other types of MGs using the proposed model has been illustrated in Table 3.2 which have good agreements with the reported value in literature [62].

Die Selection

Table 3.2 SFE of MGs based on the presented model

MGs SFE (mJ/mm2) Pt57.5Cu14.7Ni5.3P22.5 1039 Zr57Cu15.4Ni12.6Al10Nb5 997 Pd43Ni10Cu27P20 924 Au49Ag5.5Pd2.3Cu26.9Si16.3 778 Zr58.5Cu15.6Ni9.9Al10.3Nb2.8 950 La60.5Al16.3 (Cu, Ni)23.2 597

3.5.2 SFE in TPF

SFE plays a key role in TPF of MGs especially at micro and nanoscales [6].

However, this effect is not the same at all length scales. While at lower than 5 µm, this factor has significant effect on the pressure required for TPF of MGs, at larger scales it can be properly ignored in the calculations. Based on the model presented in the previous section the SFE effect on the TPF of MGs is calculated based on the following equation:

( ) P = (3-3) 32 𝜐𝜐𝜂𝜂𝑎𝑎𝑎𝑎𝑎𝑎𝐿𝐿 4𝛾𝛾 cos 𝜃𝜃 2 𝑑𝑑 − 𝑑𝑑 where P is pressure, is velocity, is apparent viscosity of MGs, is filling

𝑎𝑎𝑎𝑎𝑎𝑎 length, is die size, 𝜐𝜐 is SFE of MG𝜂𝜂 and is the wetting angle between MG𝐿𝐿 and die.

The second𝑑𝑑 term in 𝛾𝛾Eq. 3-3 is capillary 𝜃𝜃force and includes the effects of SFE and wetting angle between MGs and die on the required pressure in TPF [62]. According to the SFE calculation and wetting angle of MGs with different dies, this term has been calculated and revealed for Pt-based and Zr-based MGs on quartz in Table 3.3.

The wetting angles for Pt-based and Zr-based MGs are 145˚ and 83˚, respectively

Die Selection

[83]. For Pt-based MG due to its high wetting angle and higher SFE, the capillary pressure raises the pressure required for forming. However, for Zr-based MG this term decreases the pressure needs for forming. Although at higher length scales (1

µm) this term can be ignored, at nanometer scales its impact on forming is tangible.

Thus, based on Table 3.3 it can be concluded that for Pt-based MG the nano scale forming would be harder than Zr-based MG, because higher pressure is needed for the former.

Table 3.3 Contribution of capillary pressure on TPF of two different MGs on quartz

Capillary pressure Capillary pressure of Pt- Feature size (µm) of Zr-based MG based MG (MPa) (MPa) 100 -0.034 0.00134 10 -0.34 0.0134 1 -3.4 0.134 0.1 -34 1.34 0.01 -340 13.4

In Table 3.4 capillary pressures for Pt-based MG and two different materials, quartz and Si has been compared. The wetting angle between Pt-based MG and quartz is

145˚ and with Si is 10˚ [83]. At high die size, the effect of both dies can be ignored.

Nevertheless, it is evident that Si die is superior to quartz at lower sizes, as it reduces the required pressure for TPF. However, it should be mentioned that Si exhibit high adhesion tendency with MGs, which makes MG/Si separation very difficult.

Die Selection

Table 3.4 Effect of material and size on capillary pressure in TPF of MGs

Capillary pressure Capillary pressure Die size (µm) Quartz (MPa) Si (MPa) 100 -0.034 0.0409 10 -0.34 0.409 1 -3.4 4.09 0.1 -34 40.9 0.01 -340 409

3.6 Work of adhesion between MGs and die materials

The work of adhesion has been widely used to evaluate the adhesion degree of two adjacent materials [81, 82]. To quantitatively understand the above adhesion phenomena, a theoretical analysis on the work of adhesion between MGs and die materials will be carried out in this section. According to Dupre’s formula [149], the work of adhesion (Wad) between MGs and die materials can be expressed as;

Wad= γMG + γDie – γMG-Die (3-4) where γMG is the SFE of MG, γDie is the SFE of the die material, and γMG-Die is the

SFE of MG-Die interface formed at the molding temperature. The values of γDie can be found in Table 3.1. A macroscopic atom model has been widely used for calculating the surface and interfacial energies of crystalline materials [135].

To obtain the interfacial energy of two materials, one should calculate the interaction energies between the elements of the materials on each side of the interface [81] by using the following equation;

∆H Interface γ Interaction = S S A in C A in C CACC 2 / 3 (3-5) c0VA

Die Selection

Interface where atoms A and C belong to two different materials, ∆H A in C is the interface enthalpy for an atomic cell A fully surrounded by atomic cell C [81], the value of which can be found in [135]. Eq. (3-5) can be used for liquids, crystalline and amorphous alloys. In order to obtain the interfacial energies between La-based and

Interaction Zr-based MGs and various dies, γ for each possible reaction should be determined and add together. Take La-based MG and Si for instance, the values of

Interaction Interaction Interaction Interaction γ La in Si , γ Al in Si , γ Ni in Si and γ Cu in Si should be evaluated and added

Interface together to get γMG-Die. It is noted that the values of interface enthalpy ∆H in

[135] were calculated based on the assumption that the atomic cell A is fully surrounded by atomic cells C. This condition could be satisfied only when the materials are in liquid/gas state [81]. Considering that the molding temperatures in this study are much lower than the melting temperatures of die materials, the interface enthalpy could also be modified accordingly by homologous temperature

T/Tm.

With the calculated values of γMG and γMG-Die and the value of γDie in Table 3.1, one can get the work of adhesion for most of MG-Die pairs as summarized in Table 3.5.

The work of adhesion for PTFE was not calculated due to the lack of experimental data for the interaction between fluorine and other elements. Nevertheless, based on

Eq. 3-4, it is highly predictable that the results should be much lower than sapphire, because of the large difference in their SFE (Table 3.1). With the same die materials,

Zr-based MG has higher value of the work of adhesion than that of La-based MG, because the SFE of the former is higher. PTFE-La-based MG evaluated to possess the lowest work of adhesion, indicating it is easy to separate these two materials after

Die Selection

molding. For WC-Co also due to lack of data regarding SFE, and complexity of its structure, the work of adhesion has not been dedicated in Table 3.5. However, the calculated work of adhesion agrees very well with our experimental observation for the other dies. PTFE and sapphire had low values of the work of adhesion with two

MGs, same as the experimental observations.

Table 3.5 SFE of MGs and estimated work of adhesion in different circumstances

Surface Homologous Work of Die energy of MGs temperature adhesion Materials MGs (T/T ) (mJ/m2) (mJ/m2) m Electroless La-based MG 597 0.4 1774.62 Ni-P La-based MG Si 597 0.26 1932.3 PTFE on La-based MG 597 0.72 <<1395 copper La-based MG Sapphire 597 0.186 1395 La-based MG SiC 597 0.144 1584 Zr-based MG Si 950 0.43 2393 1892 Zr-based MG Sapphire 950 0.315

Zr-based MG SiC 950 0.24 2041

3.7 Summary

In this Chapter a number of experimental observations and modelling have been employed for die selection in TPF of MGs. It was clear that PTFE, WC-Co and sapphire had the lowest adhesion with MGs. Among them PTFE was not an appropriate die for TPF of MGs, due to its low melting point. However, sapphire

Die Selection

(Tm=2050˚C) and WC-Co (Tm=2870˚C) had desirable melting point. Both of the dies possess very good tribological properties making them suitable for TPF process. But

WC-Co had advantages in manufacturing over sapphire. The machinability and mechanical properties of WC-Co was superior to sapphire. WC-Co was also cheaper than sapphire wafer. In addition, fabrication of deep features on WC-Co is feasible; however in sapphire the depths of features were limited to 500 µm because of the wafer thickness. In conclusion, WC-Co was selected as the die for TPF in the next

Chapters.

It should also be mentioned that La-based MG was only used to enable me to consider more dies, as some of the dies were only applicable at low temperatures

(PTFE and Electroless Ni-P) and also generalize the conclusion. La-based MGs are not applicable in many applications due to low glass transition temperature and lower mechanical properties. Accordingly, the focus of this thesis in the next

Chapters is Zr-based MGs with high glass transition and mechanical properties which have potential applications in many fields rather than La-based MG with limited applications due to its low mechanical properties and temperature application.

TPF of MG on Flat Die

Chapter 4 TPF of MG on Flat Die

TPF of MG on Flat Die

After revealing the suitable die material, in this Chapter TPF of MGs at different temperatures, as the main parameter in TPF, on a flat WC-Co die is carried out. The apparent viscosity variation of MG with time at each forming temperature is calculated and studied. The structures of the MGs after TPF at different temperatures are investigated through XRD and HRTEM. In addition, the effect of TPF on electron diffraction is examined by comparing the electron diffraction patterns of the as-received MG and thermoplastically formed samples at different temperatures.

Thermal analysis investigations are also conducted by using DSC on the as-received and thermoplastically formed samples.

4.1 Experimental procedure

Zr58.5 Cu15.6Ni12.8Al10.3Nb2.8 with commercial name of LM106a was used in TPF process. The glass transition and crystallisation temperatures of this material were

390 ˚C and 497 ˚C, respectively. The amorphous structure of the as-received sample has been verified by XRD (Fig. 3.2). LM106a is one of the best glass formers among the Be-free Zr-based MG and is environmentally friendly [150]. Among the MG family, Zr-based MGs are recognized with their high SCLR temperature and desirable mechanical properties and have applications in different fields, accordingly. TPF processes were conducted at six different temperatures 450 ˚C, 460

˚C, 470 ˚C, 480 ˚C, 490 ˚C and 496 ˚C. Temperatures higher than 496 ˚C were not applicable for TPF, as the risk of crystallization in 60s was very high. At temperatures lower than 450 ˚C the high viscosity of LM106a TPF is not applicable in microforming and thus these temperatures are not studied. In all the experiments the applied load was 200 N and considering the initial diameter (5 mm) the pressure

TPF of MG on Flat Die

was held lower than 10 MPa. The loading rate changed at different temperatures, depending on viscosity, to maintain the constant load during the test This pressure ensured the Newtonian regime deformation of MG throughout the TPF tests. At high pressure the MG behaves like non-Newtonian fluid which significantly reduces the formability. Loading time was held constant around 60 s in all the tests. After experiment the sample were held inside the machine to cool down to room temperature under protection. The TPF processes were conducted on a Toshiba precision glass moulding machine (GMP-211) in the Nano and Precision

Engineering Lab at the UNSW Australia (Fig. 3.3). All the experiments are carried out under a protected environment (nitrogen) to minimise the oxidation during TPF.

After each test load displacement curve at each temperature was recorded and extracted for further analysis.

The surface roughness of the WC-Co die reached to 20 nm by grinding and polishing through nanomilling. The roughness of the WC-Co dies was measures by means of

Zygo machine (New View 700 3D optical surface profilometer) in nano and precision engineering lab at UNSW (Fig 4.1). This facility uses a scanning white light interferometry technique which is the latest one in interferometry for imaging and measuring surface topography. This machine is also equipped with an optical microscope and a digital camera with the capability of three dimensional measurements. The magnification of the used optical microscope is 10× which provide a scan size of 0.94 mm× 0.7 mm. The high speed digital camera has the resolution of 640 × 480 pixels. The roughness of the surfaces was held constant throughout the experiments in order to minimize roughness effect on the results.

TPF of MG on Flat Die

Figure 4.1 Zygo instrument in nano and precision engineering lab at UNSW

The apparent viscosity of the samples at each temperature is measured via load- displacement curve obtained from the glass moulding machine. The apparent viscosity variation with time is thoroughly explored and the mechanism of variation is revealed. After TPF, the structure of the specimens were characterized by XRD

(PANalytical X’pert) with Cu kα (λ= 0.15406 mm) with the current and voltage of

40mA and 45 kV, respectively. Moreover, HRTEM was also employed for further characterization of structure. The samples are prepared by focused ion beam (FIB) method. Because of the sensitivity of the MG structure, thin layers of gold and carbon coatings were sputtered on the MG to protect them during FIB. The SEM model XT Nova Nanolab 200 was utilized for FIB. This is a dual beam instrument, which combines high resolution ion beam and high resolution SEM. The model

Phillips CM 200 was employed for HRTEM analyses. Both structure and diffraction patterns of the samples were examined during TEM analyses.

TPF of MG on Flat Die

Thermal analyses were performed through a differential scanning calorimetry (DSC) instrument (NETZSCH 204F1 Phoenix) at a heating rate of 20 K/min to reveal the heat release and enthalpy changes of the as-received and the processed samples.

4.2 Characterisation after TPF

4.2.1 Load displacement curves

Load-displacements curves of the samples during TPF at different temperatures are shown in Figs 4.2-4.7.

0.25

0.2

0.15

0.1 Load (KN)

0.05

0 0 0.02 0.04 0.06 0.08 0.1 Displacement (mm)

Figure 4.2 Load displacement curve of the sample thermoplastically formed at 450˚C.

TPF of MG on Flat Die

0.25

0.2

0.15

0.1 Load (KN)

0.05

0 0 0.02 0.04 0.06 0.08 0.1 Displacement (mm)

Figure 4.3 Load displacement curve of the sample thermoplastically formed at 460˚C

0.25

0.2

0.15

0.1 Load (KN)

0.05

0 0 0.05 0.1 0.15 0.2 Displacement (mm)

Figure 4.4 Load displacement curve of the sample thermoplastically formed at 470˚C

TPF of MG on Flat Die

0.25

0.2

0.15

0.1 Load (KN)

0.05

0 0 0.05 0.1 0.15 0.2 0.25 0.3 Displacement (mm)

Figure 4.5 Load displacement curve of the sample thermoplastically formed at 480˚C

0.25

0.2

0.15

0.1 Load (KN)

0.05

0 0 0.2 0.4 0.6 Displacement (mm)

Figure 4.6 Load displacement curve of the sample thermoplastically formed at 490˚C

TPF of MG on Flat Die

0.25

0.2

0.15

0.1 Load (KN)

0.05

0 0 0.2 0.4 0.6 0.8 1 Displacement (mm)

Figure 4.7 Load displacement curve of the sample thermoplastically formed at 496˚C

It is clear that with increasing temperature the maximum displacement of the samples increased. The maximum displacements of the sample were 0.059 mm,

0.093 mm, 0.157 mm, 0.279 mm, 0.511 mm and 0.796 mm at 450 ˚C, 460 ˚C, 470

˚C, 480 ˚C, 490 ˚C and 496 ˚C, respectively. The first parts of the diagrams which is almost linear, is the load-displacement variation until the load reaches to 200N. After that load would be constant but the displacement continues. This is similar to creep of materials that under constant load, plastic deformation of materials continues.

This creep behaviour of MGs is the result of viscoelastic properties of these alloys that under constant load and temperature deformation continues with time.

Obviously, the lower the viscosity the higher the deformation under constant load would be. With temperature rise, viscosity of the MGs decreases significantly.

According to the load-displacement diagrams valuable information regarding the formability and apparent viscosity of MGs during TPF could be established which will be dedicated in the next sections.

TPF of MG on Flat Die

4.2.2 Formability

Qualitatively, the formability can be related to the maximum strain that MGs undergoes in their SCLR at each temperature [120]. Considering the initial dimensions of the samples and the load and displacement data given by the glass moulding machine, the strain values at each temperature were calculated. Under constant load, it is observed from Fig. 4.2-4.7 that maximum strain of the samples increases with temperature as is illustrated in Fig. 4.8. Accordingly and considering the strain-formability relationship, it is understood that the formability of LM106a, increases significantly with temperature rise. Meanwhile, considering the inverse relationship of formability with crystallisation, it also reveals that crystallization is inhibited until 496 °C. Formability is inversely corresponded with viscosity of MGs.

Accordingly, it is concluded that the apparent viscosity of LM106a decreases with increasing temperature. However, the exact value of viscosity should be calculated to have a deeper understanding of TPF ability at each temperature. In the next section the viscosity values are presented.

TPF of MG on Flat Die

0.4 0.35 0.3

0.25 0.2 0.15 Max Strain Max 0.1 0.05 0 440 450 460 470 480 490 500 Temperature (˚C)

Figure. 4.8 Maximum strains at different temperatures in TPF

4.2.3 Apparent viscosity

The theoretical temperature dependence of viscosity of an MG can be expressed via

VFT equation (Eq. (2-6)). For LM 106a, the pre-exponential factor equals to

0 5×10-5 Pa.s, which is the VFT temperature equals to 430K and D𝜂𝜂 is fragility ∗ parameter equals𝑇𝑇 to 21 [150]. Table 4.1 shows the estimated theoretical viscosity of

LM106a at the different forming temperatures calculated by VFT equation.

Table 4.1 Calculated viscosity of LM106a at different temperatures

Temperature (˚C) Viscosity (Pa.s) 450 3.39 × 109 460 1.18 × 109 470 4.43 × 108 480 1.76× 108 490 7.38× 107 496 1.15× 107

TPF of MG on Flat Die

The apparent viscosity (real viscosity in experimental condition) is different from the theoretical values which might be due to the effect of several factors including, structure, friction, oxidation and others. To calculate the apparent viscosity, it is necessary to find out whether the flow behaves Newtonian or non-Newtonian during the homogeneous flow of the experiments. This is validated by Reynold’s number in fluid mechanics:

= (4-2) 𝜌𝜌𝜌𝜌𝐷𝐷𝑒𝑒 𝑅𝑅𝑒𝑒 𝜂𝜂 7 10 by considering the η viscosity of MGs (10 to 10 Pa.s), v is velocity in µm/s, De is diameter in mm and ρ is density of MG and equals 6.5 g/cm3, one can find that Re

<< 1 and the flow behaves like Newtonian flow [151].

In Newtonian flow viscosity is independent of strain rate and mainly corresponded with temperature. Apparent viscosity of MG can be calculated via:

= (4-3) 𝜎𝜎 𝜂𝜂 3𝜀𝜀̇ In this Eq. 4-3 is apparent viscosity, is the stress and is the strain rate during

TPF. is calculated𝜂𝜂 according to the force𝜎𝜎 and displacement𝜀𝜀̇ of the sample at each temperature𝜎𝜎 . For calculation of strain rate, strain at each time is calculated based on the displacement and then divided by the corresponded time at each moment. Under non-Newtonian condition this ratio is not applicable. Apparent viscosity is a term that used for determining the viscosity of MG under real condition and might be close or different with theoretical viscosity. Because viscometer is not applicable for these alloys, researchers use this term to establish the fluid behaviour in real working condition. Figs. 4.9-4.14 show the variation of apparent viscosity of LM106a with

TPF of MG on Flat Die

time during TPF. It is evident that viscosity of all the samples increases with time. At lower temperatures before 480 °C, viscosity increased during the test. However from

480 °C the slopes decreases and apparent viscosity reaches to a plateau at 490 °C and

496 °C.

6000

5000

4000

3000

2000

Apparent viscosity (MPa.s) viscosity Apparent 1000

0 0 10 20 30 40 50 60 time (s)

Figure 4.9 Viscosity variations with time at 450˚C

TPF of MG on Flat Die

4000

3500 3000 2500 2000 1500 1000

Apparent viscosity (MPa.s) viscosity Apparent 500 0 0 10 20 30 40 50 60 time (s)

Figure 4.10 Viscosity change with time at 460˚C

2500

2000

1500

1000

500 Apparent viscosity (MPa.s) viscosity Apparent

0 0 10 20 30 40 50 60 time (s)

Figure 4.11 Viscosity change with time at 470˚C

TPF of MG on Flat Die

1400

1200

1000

800

600

400

Apparent Viscosity (MPa.s) Viscosity Apparent 200

0 0 10 20 30 40 50 60 time (s)

Figure 4.12 Viscosity change with time at 480˚C

600

500

400

300

200

100

Apparent viscosity (MPa.s) viscosity Apparent 0 0 10 20 30 40 50 60 time (s)

Figure 4.13 Viscosity change with time at 490˚C

TPF of MG on Flat Die

350

300

250

200

150

100

Apparent viscosity (MPa.s) viscosity Apparent 50

0 0 10 20 30 40 50 60 time (s)

Figure 4.14 Viscosity change with time at 496˚C

The variation of viscosity with time can provide us valuable information regarding the mechanism of plastic deformation. Spaepen [101] studied the viscosity variation of a Pd-based MG. The viscosity variation of the Pd-based MG has been illustrated in Fig. 4.15 which is similar to the viscosity trend at lower temperatures in this thesis.

TPF of MG on Flat Die

Figure 4.15 Variation of viscosity with time of Pd41Ni10Cu29P20 at 555 K and 15 MPa due to structural relaxation [101]

The viscosity time variation for MGs is unique depending on the temperatures. It is shown that at glassy state (lower than glass transition temperature), where MG is far from the equilibrium, viscosity variation with time is linear [101, 152]. However in

SCLR the viscosity variation becomes nonlinear. In SCLR, when MG approaches the equilibrium state the viscosity reaches to a plateau and its variation becomes zero as shown in Figs 4.13 and 4.14. This can be explained by free volume model.

According to free volume model, at each temperature there is an equilibrium free volume concentration. By reaching to this level, free volume concentration becomes constant approximately. Meanwhile, at a given temperature, viscosity and free volume concentrations are inversely correlated ( = ). Therefore, increasing 1 𝜂𝜂 − 𝑐𝑐𝑓𝑓 viscosity of MGs indicates free volume annihilation. It is believed that structural

TPF of MG on Flat Die

relaxation is the primary reason behind free volume annihilation [100, 152]. Since structural relaxation is strongly temperature dependent, its rate increases substantially with temperature. As a result under a fixed forming time (60s), free volume annihilation would be faster by increasing temperatures. At higher temperatures 496 ˚C and 490 ˚C, where MGs approaches equilibrium, free volume concentration reaches to its minimum value and would not change. Accordingly, viscosity reaches to a plateau. At lower temperatures, however, which are far from equilibrium, viscosity does not reach to its equilibrium value during the experiments

(60 s). This phenomenon and its effect on TEM diffraction pattern mechanical properties will be further explained in section 4.1.4 of this chapter and chapter 7, respectively.

4.2.4 XRD analysis

The XRD result of the samples thermoplastically formed at 490˚C and 496˚C is revealed in Figs. 4.16 and 4.17. It is evident that the sample shows typical halo shape pattern of amorphous materials without any peak. As amorphous structure of MGs verified at higher temperature in Figs 4.16 and 4.17, it would be evident that the samples thermoplastically formed at lower temperature are also possess amorphous after TPF.

TPF of MG on Flat Die

6000

5000

4000

3000

Intensity 2000

1000

0 20 30 40 50 60 70 2θ (degree)

Figure 4.16 XRD result of the sample thermoplastically formed at 490˚C

3500 3000 2500

2000 1500 Intensity 1000 500 0 20 30 40 50 60 70 2θ (degree)

Figure 4.17 XRD result of the sample thermoplastically at 496˚C

4.2.5 HRTEM analysis

In order to prepare samples for HRTEM analyses, FIB technique was used. Before

FIB and for the protection of the surface, the samples were coated with gold and

TPF of MG on Flat Die

carbon layers. Figs. 4.18-4.20 show the SEM images of the samples prepared for

HRTEM analyses.

Figure 4.18 SEM image of the as received sample prepared by FIB for TEM analysis

Figure 4.19 SEM image of the thermoplastically formed sample at 450 ˚C sample prepared by FIB for TEM analysis 97

TPF of MG on Flat Die

Figure 4.20 SEM image of the thermoplastically formed sample at 496 ˚C prepared by FIB for TEM analysis

The structure and diffraction patterns of the as-received MG, the sample thermoplastically formed at 450 ˚C and 496 ˚C are illustrated in Figs. 4.21-4.23, respectively. The amorphous structure without any short to medium range order can be observed for the first two specimens.

The diffraction patterns of the samples are also typical halo ring shape of amorphous materials without any spot or ring. These patterns confirm the amorphous structure of the all samples after TPF processes. However, it is interesting to note that the halo ring diameter of the as-received and thermoplastically formed samples do not have identical size. The ring diameters for the as-received sample and thermoplastically formed at 450 °C and 496 °C samples are 7.96 1/nm, 8.5 1/nm and 8.58 1/nm, respectively. After TPF, the ring diameter increases; yet, ring diameter increases with temperature rise. It is known that the ring diameter is inversely corresponded with interatomic spacing in MGs [32]. It means that when the ring diameter is

TPF of MG on Flat Die

smaller, the interatomic spacing becomes larger. Interatomic spacing also directly attributed to free volume concentration in MGs [31, 32]. Thus, MGs with lower interatomic spacing contain lower free volume concentration. According to greater diameter size of the diffraction pattern of the thermoplastically formed samples compared with the as-received MG (Figs 4.21b-4.23b), it is concluded that after TPF free volume concentration of the samples becomes smaller; yet free volume further decreases with forming temperature. Thus both TEM analyses along with apparent viscosity variation with time verify free volume annihilation during TPF. The effect of this phenomenon on mechanical properties of MGs will be thoroughly studied in

Chapter 7.

Figure 4.21 Structure and diffraction pattern of the as-received material

TPF of MG on Flat Die

Figure 4.22 Structure and diffraction pattern of the sample thermoplastically formed at 450˚C

Figure 4.23 Structure and diffraction pattern of the sample thermoplastically formed at 496 ˚C

4.2.6 DSC analyses

As mentioned earlier it is believed that structural relaxation is the main reason of free volume annihilation. In this section and in order to investigate the effect of TPF on the structural relaxation and free volume annihilation, DSC tests were performed on the as-received sample and thermoplastically formed MGs at 450 ˚C and 496 ˚C. 100

TPF of MG on Flat Die

Figs. 4.24-4.26 show the corresponded DSC results of the MGs. It is clear that the heat flow of the samples thermoplastically formed is smaller than the value for as- received MG. The heat release decreases from 28.7 J/g for the as-received MG to

26.7 J/g and 20.9 J/g for the sample formed at 450 ˚C and 496 ˚C, respectively. This clearly demonstrates that structural relaxation occurs as a result of TPF. In addition it is realized that higher forming temperature accelerates the structural relaxation. It is also found that Tg of the sample remain almost the same around 390 ˚C and TPF did not affect the glass transition temperature of the MG. Structural relaxation correlated directly with free volume reduction of MGs [153, 154]. It is thus understood from the relaxation enthalpy reduction that free volume annihilation occurs during TPF as a result of structural relaxation.

Figure 4.24 DSC result of the as-received MG

101

TPF of MG on Flat Die

Figure 4.25 DSC result of the thermoplastically formed MG at 450˚C

Figure 4.26 DSC result of the thermoplastically formed MG at 496˚C

102

TPF of MG on Flat Die

4.3 Summary

In this Chapter a number of experiments are conducted at different temperatures within SCLR of LM106a on flat WC-Co die. It was found that temperature plays a key role in formability, apparent viscosity, structure and structural relaxation of the

MG in TPF. With increasing temperatures formability of MG increased significantly.

Apparent viscosity of the samples decreased from 5 × 109 at 450 ˚C Pa.s to 3× 108

Pa.s at 496 ˚C. In addition, viscosity time variation study revealed that free volume annihilation occurs at different temperature causing viscosity increase with time. At higher temperatures (496 ˚C and 490 ˚C), apparent viscosity approached to its equilibrium values and reached to a plateau. While at lower temperature the viscosity was far from equilibrium and increased with time. The amorphous structure of MG after forming was verified by means of XRD. HRTEM analyses of the sample before and after TPF also demonstrated the amorphous structure of samples after forming. It was found that diffraction patterns were typical halo ring shape in all the samples.

But the size of the ring diameter increased after TPF. The ring diameter after TPF even became larger by increasing the forming temperature. Considering the inverse relationship between ring diameter in diffraction pattern and interatomic spacing, it was concluded that free volume annihilation happened during TPF. In addition, the

DSC analysis of the samples confirms the occurrence of structural relaxation during

TPF which leads to free volume annihilation.

103

Micro Forming of MG Using a Microchannelled Die

Chapter 5 Micro Forming of MG Using a Microchannelled

Die

104

Micro Forming of MG Using a Microchannelled Die

Miniaturization of products has become irresistible trend due to the rapid development of microelectronics and information technology. Microforming process is the extension of traditional plastic forming technology. Due to superplasticity and lack of grain boundary in structure, MGs can be easily shaped in micro scales in

SCLR. Micro formability of amorphous alloys depends strongly on the process parameters such as the load, processing time and temperature, as well as the shape of the micropattern.

After discovering the die material and proper forming parameters in Chapters 3 and

4, in this Chapter TPF is conducted at different temperatures by using a microchannelled die. After forming the structure and apparent viscosity of the products are examined.

5.1 Fabrication and characterisation of die

As discussed in Chapter 3, WC-Co was selected as the die materials. This material would be ideal for microforming of MGs because of its low adhesion status with

MGs, high melting point, good mechanical properties and very close thermal expansion coefficient with LM106a which facilitates the separation and minimizes thermal stress within the products. Moreover, the machinability of this material is superior to the other candidates like PTFE or sapphire enabling to fabricate different features with excellent roughness on WC-Co.

Due to the high hardness of WC-Co only diamond tools were applicable for machining the surface. Accordingly a polycrystalline diamond square end mill

105

Micro Forming of MG Using a Microchannelled Die

(PCDSE) is purchased from NS-tool company, Japan and employed for machining.

Fig. 5.1 shows the specification of the employed tool. The tool has a square shape end with the width size of 100 µm.

Figure 5.1 Schematic and specifications of PCDSE

To fabricate the microchannel, the Nano 350 FG 5 Axis Ultra Precision Lathe facility installed in the laboratory of precision and nano processing technologies at

UNSW was employed (Fig. 5.2). The milling mode of the instrument was utilized in this process. The machining parameters were chosen based on company’s recommendation. Accordingly, the spindle speed was 40000 min-1 and the feed rate was 0.125 mm/min. Before milling process, the surface of the WC-Co die was ground to mirror surface. After grinding, a channel shape die with 100 µm of width and 58 µm of depth is fabricated on WC-Co die.

106

Micro Forming of MG Using a Microchannelled Die

Figure 5.2 Nano 350 FG 5 Axis Ultra Precision Lathe

Figures 5.3 and 5.4 illustrate the optical microscope images with different magnifications of the fabricated microchannelled WC-Co die.

Figure 5.3 Microchannelled WC-Co die

107

Micro Forming of MG Using a Microchannelled Die

Figure 5.4 Fabricated WC-Co microchannelled at higher magnification

The profile specifications and surface roughness of the microchannelled die are investigated by Zygo machine and the results are revealed in Figs. 5.5 and 5.6. It is evident that the microchannel has 100 µm width and 57.576 µm depth. The surface roughness of the microchannel is around 20 nm.

Figure 5.5 Zygo image of the channel profile on the WC-Co

108

Micro Forming of MG Using a Microchannelled Die

Figure 5.6 Roughness of the channel surface

5.2 Experimental set up

After die fabrication TPF process was set up by choosing a proper load, a loading time and a temperature. LM106a was selected as the raw materials due to its great

GFA and wide range of SCLR. The TPF was conducted at different temperature

450˚C, 460˚C, 470˚C, 480˚C, 490˚C and 496˚C. The force was set to 200 N and forming time was around 60s. The TPF processes were conducted on a Toshiba precision glass moulding machine (GMP-211) in the Nano and Precision

Engineering Lab at the UNSW Australia. The samples during forming process were under the protection of nitrogen gas in order to minimize the oxidation effect. After

TPF, the samples were held in the machine so as to be cooled.

109

Micro Forming of MG Using a Microchannelled Die

After TPF the structure of MGs are characterised by optical microscope and Zygo interference microscopy. The apparent viscosity of the microribs was calculated based on their filling lengths. The mechanical properties are also measured by microhardness tester (Chapter 7).

5.3 Characterisation of the microribs

After TPF processes the microribs at different temperatures were characterised by using optical microscope and Zygo machine.

5.3.1 Microscopy analysis

Figures 5.7–5.12 illustrate the microchannels fabricated at temperatures between 450

˚C to 496 ˚C. Oxidation layers are observed on the microchannel at all temperatures, though the protection is carried out during moulding. This is because of high tendency of the elements such as Zr, Al, Nb to react with oxygen.

Figure 5.7 Microrib thermoplastically formed at 450˚C

110

Micro Forming of MG Using a Microchannelled Die

Figure 5.8 Microrib thermoplastically formed at 460˚C

Figure 5.9 Microrib thermoplastically formed at 470˚C

111

Micro Forming of MG Using a Microchannelled Die

Figure 5.10 Microrib thermoplastically formed at 480˚C

Figure 5.11 Microrib thermoplastically formed at 490˚C

112

Micro Forming of MG Using a Microchannelled Die

Figure 5.12 Microrib thermoplastically formed at 496˚C

5.3.2 Roughness analysis

Roughness of the all microribs is also investigated by using Zygo machine. The roughness of the microribs was around 50 nm and it did not show any tangible difference with temperature.

5.3.3 Apparent viscosity

Apparent viscosity is an effective factor that quantifies formability of MG. This term includes all the internal and external parameters in experiments and therefore would be useful in designing and controlling the TPF processing. The Reynold’s number

(Re) is a widespread factor in fluid mechanics which is pivotal in determining the flow bahviour (Eq. 4-2) [151, 155]:

113

Micro Forming of MG Using a Microchannelled Die

In microforming of MGs where the viscosity is a large value (106 to109 Pa.s) and the dimensions and velocity are in micrometer and micrometer per second respectively, the Reynold’s number would be <<1 and the flow can be considered as Newtonian

[120, 151].

De in Eq. 5-1 is equivalent hydraulic diameter and according to fluid mechanics is calculated [156]:

De=4A/g (5-1)

Where A is the area and g is the perimeter of a die shape. De for some popular shapes can be found in [151].

As illustrated in Fig. 5.13, by assuming an MG as a fluid that fills a channel, the resulting force acting on a controlled volume (CV) with a radius of r and length dx equals to:

+ . + . 2 . = 0 2 𝛿𝛿𝛿𝛿 2 𝑝𝑝𝜋𝜋𝜋𝜋 − �𝑝𝑝 𝑑𝑑𝑑𝑑� 𝜋𝜋𝑟𝑟 𝜏𝜏 𝜋𝜋𝜋𝜋 𝑑𝑑𝑑𝑑 𝛿𝛿𝛿𝛿 (5-2)

= (5-3) 𝑑𝑑𝑑𝑑 2𝜏𝜏 𝑑𝑑𝑑𝑑 𝑟𝑟

114

Micro Forming of MG Using a Microchannelled Die

Figure 5.13 The schematic tube for force analysis

Where p is the pressure at the cross section area of CV and τ is Newtonian viscous stress acting on the surface of perimeter of CV. According to Newton’s law, viscosity is related to shear stress:

= (5-4) 𝑑𝑑𝑑𝑑 𝜏𝜏 𝜂𝜂 𝑑𝑑𝑑𝑑 where v is the velocity of the fluid at distance r from the central axis of the tube.

Substituting equation τ from equation 5-5 in Eq. 5-4 we have:

= . (5-5) 𝑑𝑑𝑑𝑑 1 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 2𝜂𝜂 𝑑𝑑𝑑𝑑 𝑟𝑟 By integrating the equation, we have

= . + (5-6) 1 𝑑𝑑𝑑𝑑 2 𝑣𝑣 4𝜂𝜂 𝑑𝑑𝑑𝑑 𝑟𝑟 𝐶𝐶 C is an integral constant. According to no-slip boundary condition for viscous flow, v=0 at when r=De/2. By calculating C and putting in the Eq. 5-7 we have:

= . ( ) (5-7) 2 1 𝑑𝑑𝑑𝑑 2 𝐷𝐷𝐷𝐷 𝑣𝑣 4𝜂𝜂 𝑑𝑑𝑑𝑑 𝑟𝑟 − 4 115

Micro Forming of MG Using a Microchannelled Die

By integrating the velocity distribution over the whole area, the volume flow rate (Q) of the MG is established:

Q= 4 (5-8) 𝜋𝜋𝐷𝐷𝑒𝑒 𝑑𝑑𝑑𝑑 128𝜂𝜂 𝑑𝑑𝑑𝑑

For the filling length L and applied pressure P and filling time t we have:

= 2 4 (5-9) 𝜋𝜋𝐿𝐿𝐷𝐷𝑒𝑒 𝜋𝜋𝐷𝐷𝑒𝑒 𝑃𝑃 4𝑡𝑡 128𝜂𝜂 𝐿𝐿

L= 2 (5-10) 𝑡𝑡𝐷𝐷𝑒𝑒 𝑃𝑃 � 32𝜂𝜂

In TPF and based on Eq. 5-10 we are able to calculate apparent viscosity (ηapp):

ηapp= ( )( ) (5-11) 𝑡𝑡 𝐷𝐷𝑒𝑒 2 32 𝑃𝑃 𝐿𝐿 Under some circumstances, and as a result of the effect of SFE and capillary force on the apparent viscosity a new term will be added in to the previous equation [62, 123]

ηapp= ( + )( ) (5-12) 𝑡𝑡 4𝛾𝛾 cos 𝜃𝜃 𝐷𝐷𝑒𝑒 2 32 𝑃𝑃 𝑑𝑑 𝐿𝐿 P is calculated based on the applied load and final diameter of sample measured by calliper after test. Under severe oxidation condition, apparent viscosity is affected by oxidation and a term should be added to Eq. 5-12. Both capillary and oxidation effects on the apparent viscosity will be evaluated for the microchannelled die.

In order to calculate the apparent viscosity of the samples at each temperature, the filling lengths should be measured first. The maximum filling lengths of the MG at

116

Micro Forming of MG Using a Microchannelled Die

each temperature is measured by Zygo and revealed in Fig. 5.14. It is clear that the maximum filling depth of the sample after 480 ˚C reached to the bottom of the microchannel. For the higher temperatures where the fluid touched the bottom of the die, apparent viscosity could not be calculated. However for lower temperature (450

˚C, 460 ˚C and 470 ˚C) the apparent viscosity is calculated based on the filling length, pressure and die size.

70 60 50 40 30 20

(µm) Filling length 10 0 440 450 460 470 480 490 500 Temperature (˚C)

Figure 5 .14 Filling length variations with temperature for microribs

According to Eq. 5-2 the equivalent hydraulic diameter of the microchannelled die can be expressed as:

De= (5-13) 2𝑊𝑊𝐿𝐿𝑑𝑑𝑑𝑑𝑑𝑑 𝑊𝑊+𝐿𝐿𝑑𝑑𝑑𝑑𝑑𝑑 where W and Ldie are the width and length of the fabricated channel, respectively.

The calculated De for the microchannel is around 196 µm.

117

Micro Forming of MG Using a Microchannelled Die

With aid of Eq. 5-12, the apparent viscosity at different temperatures is calculated and illustrated in Fig. 5.15. It is obvious that the apparent viscosity is decreasing with temperature. The apparent viscosity decreased from around 6 × 1010 Pa.s at 450

˚C to 1.4 × 109 Pa.s at 470 ˚C. At higher temperatures, because the fluid reached the bottom of the die, apparent viscosity could not be calculated.

1.00E+11 1.00E+10 1.00E+09 1.00E+08 1.00E+07 1.00E+06 1.00E+05 1.00E+04 1.00E+03 1.00E+02 1.00E+01

Log Apparent Apparent Log visvosity (Pa.s) 1.00E+00 445 450 455 460 465 470 475 Temperature (˚C)

Figure 5 .15 Apparent viscosity variations with temperature for TPF of LM106a using a microchannelled die

The effects of capillary pressure and oxidation on filling length of the microribs are investigated in the following paragraphs. Capillary force is calculated via [62]:

Capillary pressure= (5-14) 4𝛾𝛾 cos 𝜃𝜃 𝐷𝐷𝑒𝑒 where is SFE of LM106a, is wetting angle between liquid LM106a and die and

is equivalent𝛾𝛾 hydraulic diameter𝜃𝜃 of the microchannel. SFE of LM106a is around

𝑒𝑒 950𝐷𝐷 mJ/m2 (calculated in chapter 3), of around 200 µm, and considering the

𝐷𝐷𝑒𝑒 118

Micro Forming of MG Using a Microchannelled Die

complete wetting condition ( = 0˚), the maximum capillary pressure is 0.016 MPa.

However, the wetting angle between𝜃𝜃 WC-Co and liquid LM106a is definitely more than 0˚ and this factor even reduces the capillary pressure to lower than 0.016 MPa.

Thus, the capillary pressure in TPF of the Microchannelled die would be ignored.

However for smaller dies, in nanometric scales, capillary pressure would be significant.

The oxidation effect on the TPF of MGs for a circular shape die has been calculated

[123]. By replacing the equivalent hydraulic diameter in to the equation, the change of pressure as a result of oxide layer for the LM106a microribs is proposed:

P0= ( ) (5-15) 16 𝜎𝜎𝑐𝑐 ℎ 2 2 3 √1−𝜐𝜐+𝜐𝜐 𝐷𝐷𝑒𝑒

In this equation P0 is the pressure required for breaking the oxide layer, is the

𝑐𝑐 strength of oxide layer, is Poisson ratio of MG, h is the oxidation layer thickness𝜎𝜎 and is the hydraulic𝜐𝜐 equivalent diameter of the die. Accordingly, with an

𝑒𝑒 exaggerated𝐷𝐷 oxide layer thickness of 1 µm, of around 1 GPa and Poisson ratio of

𝑐𝑐 0.3, the maximum required pressure for breaking𝜎𝜎 the oxide layer is around 0.0156

MPa. Considering the exaggerated values for the parameters, this pressure would be even smaller than the mentioned magnitude. However for smaller dies the oxidation pressure would be significant. For instance for a microchannelled with a 1 µm width the oxidation pressure by using the same parameters would be around 60 MPa. As it is clear for lower than 1 µm, the oxidation status of MGs would be extremely influential in TPF of LM106a. For noble MGs like Pd-based MG this factor would be much smaller than active MG. For Zr-based MG this factor is large and inhibits

119

Micro Forming of MG Using a Microchannelled Die

flowing of MGs in to the die. Thus the effect of oxide layer in the filling length and apparent viscosity can be ignored appropriately.

5.4 Summary

In this Chapter microforming of LM106a at different temperatures and using a microchannelled die with a 100 µm width and 58 µm depth was successfully conducted. Microribs products at the whole temperature could be easily separated from the WC-Co die. At temperatures 450 ˚C, 460 ˚C and 470 ˚C, the maximum flow depth of Zr-based MG, were 7 µm, 21 µm and 42 µm, respectively. However at higher temperatures the filling length was constant and around 57.5 µm, as the MG reached to the bottom of the die and could not flow further. After manufacturing the surface oxidation and roughness of the microchannel were analysed by means of optical microscope as well as Zygo. The surface roughness of the products was around 50 nm and did not change with temperature. According to the maximum depth of the microribs, the apparent viscosity was calculated. The effects of the capillary pressure and the pressure required for breaking oxide layer were calculated quantitatively. It was demonstrated that both of these factors had trivial effects on

TPF of LM106a by using a microchannelled die. In this Chapter due to depth limitation, we were not able to characterise the apparent viscosity at higher temperatures. Accordingly, in the next Chapter a deep microholed die is employed to characterise the apparent viscosity at the all forming temperatures. In addition, by conducting TPF on two identical dies, we are able to investigate die effect on the mechanical properties of the products.

120

Micro Forming of MG Using a Microholed Die

Chapter 6 Micro Forming of MG Using a Microholed Die

121

Micro Forming of MG Using a Microholed Die

In this Chapter microforming of a Zr-based MG is investigated by using a deep microholed WC-Co dies at different temperatures. After TPF the roughness, structure, filling lengths, apparent viscosity and mechanical properties (Chapter 7) of the products will be investigated.

6.1 Fabrication and characterisation of the die

As shown in Chapter 3, WC-Co is used as the die material because of its desirable mechanical properties, close thermal expansion coefficient with Zr-based MG in supercooled state, good adhesion resistance with MGs, high melting point and desirable machinability. Due to the high hardness of WC-Co, diamond tool is applicable for machining of this material. A diamond drilling tool with 500 µm diameter and 3 mm effective length was purchased and employed for making the microholed die. Figs 6.1 and 6.2 show the surface of the diamond tool at different magnifications. The diamond particles are obvious in the microscopic image of the tool. The roughness of the diamond tool was investigated by means of Zygo and was around 700 nm.

122

Micro Forming of MG Using a Microholed Die

Figure 6 .1 Diamond tool employed for manufacturing of the microholed die

Figure 6 .2 Diamond tool employed for manufacturing of the microholed die

In order to fabricate the die, the Nano 350 FG 5 Axis Ultra Precision Lathe facility installed in the laboratory of precision and nano processing technologies at UNSW was employed (Fig. 5.2). Before making the hole, the surface of the die is ground to

123

Micro Forming of MG Using a Microholed Die

20 nm by grinding mode of the nanomachining facility. Drilling mode is used for fabricating the microhole on the die. The drilling parameters for microhole fabrication were rotation speed of 40000 rpm and feed rate of 0.125 mm/min. The microhole is made at the centre of the die. Fig. 6.3 shows the microholed WC-Co utilized in this Chapter.

Figure 6 .3 Microholed WC-Co die

6.2 Experimental set up

TPF of MGs is conducted at 450 ˚C, 460 ˚C, 470 ˚C, 480 ˚C, 490 ˚C and 496 ˚C under constant load and time by using the microholed WC-Co die to reveal the die feature effect on TPF. The diameter of the hole is around 500 µm and its depth is

124

Micro Forming of MG Using a Microholed Die

around 3 mm. At temperatures lower than 450 ˚C, MG could not flow in to the die under the applied pressure. Because of the higher viscosity of MG at lower temperatures, the applied pressure was unable to overcome the viscous pressure of the MG and higher pressure was required accordingly. The applied load at all the experiments was 200 N. This is the lowest allowable load of the machine.

Considering the diameter of the MG which was 5 mm, the maximum applied pressure was around 10 MPa. This pressure provided a laminar flow during the experiments. The loading time was 60s. The highest forming temperature was very close to the crystallisation temperature (497 ˚C) of LM106a which is because of high

GFA of this alloy. The TPF processes were conducted by using a Toshiba glass moulding machine at Precision and nano processing technologies lab at the UNSW

(Fig. 3.3). All the tests were carried out under the protection of nitrogen gas to minimise the oxidation. After separating the samples from die, the final diameters of the samples and the flowing depth of the microrods were measured for evaluation of the apparent viscosity at each temperature. The reported maximum filling lengths and diameters were the average of 5 measurements. The structures of the microrods were also characterized by using HRTEM. Furthermore, the hardness of the microrods at each temperature were characterised and will be dedicated in Chapter 7.

6.3 Characterisation after TPF

Microforming was successfully conducted and microrods with different length scales were fabricated. Fig. 6.4 shows the microrod fabricated at 496 ˚C. Due to desirable adhesion resistance of WC-Co die, MGs could be simply separated from the die.

125

Micro Forming of MG Using a Microholed Die

This is an important achievement, as the die could be utilized for several tests and it did not need to be dissolved in a chemical.

Figure 6 .4 Microrod fabricated at 496 ˚C by microforming

6.3.1 Microscopy analysis

The fabricated microrods are analyzed by means of an optical microscope model

Nikon DS-Ri2 which is shown in Fig. 6.5. Figs 6.6-6.9 illustrate the side views of the microscopic pictures of the fabricated microrods at 490 ˚C and 496 ˚C, respectively. At lower temperature because of the small lengths of the microrods, optical microscope were not applicable for capturing picture.

126

Micro Forming of MG Using a Microholed Die

Figure 6 .5 Nikon DS-Ri2 optical microscope

Figure 6 .6 Optical microscope analysis of the side view of the microrods fabricated at 490˚C

127

Micro Forming of MG Using a Microholed Die

Figure 6 .7 Optical microscope analysis of the side view of the microrods fabricated at 496˚C

Figure 6 .8 Optical microscope analysis of the side view microrods fabricated at 496˚C at higher magnification

128

Micro Forming of MG Using a Microholed Die

Figure 6 .9 Top view of the microrod fabricated at 496˚C

It is clear that a thin oxide layer covers the surface of the microrods. Although TPF processes are conducted under the protection of nitrogen, it seems that oxidation is inevitable which is due to the high oxidation tendency of Zr. Oxiation happens in all the microrods and temperature did not exhibit any major influence on the oxidtion status of the microrods.

6.3.2 Roughness analysis

Figure 6.10 shows the surface roughness of the microrod fabricated at 496 ˚C. The roughness of the surface of the microrods was around 700 nm. Roughness was similar for all the microrods indicating trivial effect of temperature on the roughness and surface quality of the products.

129

Micro Forming of MG Using a Microholed Die

Figure 6 .10 Surface roughness of the microrod fabricated at 496 ˚C

6.3.3 Apparent viscosity

The lengths of the microrods are measured by means of Zygo machine and calliper.

For low temperature 450 ˚C, 460 ˚C and 470 ˚C, Zygo was used for improving the accuracy of the measurements. However, at higher temperatures the lengths were measured by calliper. The measurements were repeated 5 times and the average was reported as the microrod length as shown in Fig 6.11. The minimum length was at

450 ˚C with around 3 µm and the maximum was at 496 ˚C with around 520 µm. At lower temperature LM106a could not flow in to the die under the employed pressure.

It is clear that the length of the microrod is increasing with temperature rise. This indicates that crystallisation has been inhibited during the test, since crystallisation significantly reduces the formability of MGs [120]. However in the experiments, owing to the significant growth of microrods lengths (more than 170 times in 46 ˚C), one can conclude that crystallisation is inhibited during TPF. This will be further verified by TEM analyses.

130

Micro Forming of MG Using a Microholed Die

600

500

400

300

200

(µm) length Microrod 100

0 440 450 460 470 480 490 500 Temperature (˚C)

Figure 6 .11 Microrod length variations with temperature

As shown and proved in Chapter 5, filling length is inversely corresponded to apparent viscosity and therefore it is concluded that the apparent viscosity of MG is decreasing substantially with temperature rise.

In order to calculate the apparent viscosity Eq. 5-11 is used. According to this, the apparent viscosity at each temperature is presented in Fig. 6.12. It is clear that the apparent viscosity drops significantly with increasing temperature from 450 ˚C to

496 ˚C. At lower temperatures MG could not flow inside the die and based on the

Eq. 6.1, ηapp approached infinity. The maximum apparent viscosity which enables the supercooled liquid to flow in to a 500 µm die under 10 MPa was 6.6 × 1010 Pa.s at 450 ˚C. It should be mentioned that the pressure of around 10 MPa is employed to compare the formability of LM106a with polymers, because in polymers almost any micro sized-features can be moulded by pressure less than 10 MPa [157]. The lowest achievable viscosity in TPF of LM106a was 1.6× 107 Pa.s approximately obtained at 131

Micro Forming of MG Using a Microholed Die

496 ˚C. Though this value might change slightly by changing the parameters, it would be an appropriate indicator showing the approximate minimum attainable viscosity in the microforming of LM106a. Moreover, Eq. 6-1 suggests that under constant pressure, forming time and temperature, L/De ratio should remain constant.

The independency of apparent viscosity to die size under same TPF condition in microscales has been experimentally verified in [6, 151]. The comparison of apparent viscosity of microrods and microribs at temperatures 450˚C, 460˚C and

470˚C show that the values are close to each other e.g. at 450˚C 5.9×1010 (microrib)

Vs 6.6×1010 (microrod). However further analyses at higher temperatures are required to obtain a solid conclusion.

1E+11 1E+10 1E+09 1E+08 1E+07 1E+06 1E+05 1E+04 1E+03 1E+02 1E+01 1E+00 Log Apparent Log Apparent viscosity (Pa.s) 440 460 480 500 Temperature (˚C)

Figure 6 .12 Apparent viscosity variation at different temperature in TPF

The apparent viscosity reported here includes all the external and internal factors such as capillary pressure, viscous pressure, oxidation and friction. Accordingly, the effects of external parameters would be investigated in the next paragraphs.

132

Micro Forming of MG Using a Microholed Die

In TPF applied pressure should overcome the resistance originating from viscosity of liquid alloy. In this case the supercooled metal is able to flow in to the dies. Except viscous pressure, generally, two other factors inhibiting the MG flow in to a cavity, oxide layer and capillary pressure [158].

LM106a is an active alloy and can be easily oxidised as shown in Figs. 6.9 and 6.10.

Before being able to flow in to the hole, LM106a must break the oxide layer on its surface. The required pressure for breaking the oxide layer is proposed [123]:

P0= ( ) (6-2) 16 𝜎𝜎𝑐𝑐 ℎ 2 2 3 √1−𝜐𝜐+𝜐𝜐 𝐷𝐷𝑒𝑒 With assumption of an exaggerated oxide layer by the thickness of 1 µm, of

𝑐𝑐 around 1 GPa and Poisson ratio of 0.3, the required pressure for breaking oxide 𝜎𝜎layer is 0.024 MPa. This value is very small compared with the applied pressure 10 MPa.

Thus, the effect of oxide layer in the filling lengths and apparent viscosity can be ignored appropriately. By decreasing the die size the effect of this factor would be important. For instance, in a 1µm diameter die, the required pressure for breaking oxide layer is more than 4 GPa for a highly oxidized MG.

The capillary pressure is another factor which resists in front of materials flow in to the die. The term formulated as:

Capillary pressure= (6-3) 4𝛾𝛾 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃 𝐷𝐷𝑒𝑒

133

Micro Forming of MG Using a Microholed Die

Considering the complete wetting among the supercooled metal and the die ( =0˚) and the SFE of around 950 mJ/m2 and of 500 µm, the maximum capillary𝜃𝜃

𝑒𝑒 pressure would be 0.0076 MPa. The real capillary𝐷𝐷 pressure value should be even less than the reported one, as we assume the complete wetting which is not achievable in

LM106a/WC-Co system. As a result, this value is extremely small compared with 10

MPa and can be properly ignored. However at lower sizes the value of capillary pressure would be significant. For instance at 1 µm, 100 nm and 10 nm, the capillary pressure would be 3.8, 38 and 380 MPa, respectively. Therefore viscosity resistance takes the pivotal part in front of micromolding of MG. In Newtonian fluids, the viscosity (viscous pressure) is independent of strain rate. Accordingly the minimum attainable viscosity of LM106a in TPF can be considered 10 7 Pa.s, approximately.

As mentioned earlier we were not able to fabricate microrod at lower than 450 ˚C by pressure of around 10 MPa. Thus compared with TPF of polymers, larger pressure is required for MGs because of the stronger nature of fluid of amorphous metals. The calculated viscosity of LM106a by using VFT equation at 496 ˚C is 1.15 ×10 7 Pa.s which is very close to the experimental value for apparent viscosity. However by decreasing temperature the difference between the theoretical and experimental values will be larger which might be due to lack of accuracy of VFT equation at lower temperatures and/or the stronger impact of external parameter such as friction on the experimental results.

134

Micro Forming of MG Using a Microholed Die

6.3.4 Structural analysis

Figures 6.13-6.15 show the HRTEM analyses and diffraction patterns of the as- received MG as well as the microrods manufactured at 450 ˚C and 496 ˚C, respectively. It is clear that the as-received material has an amorphous structure without any ordering. TPF of LM106a at 450 ˚C also does not induce any ordering in the structure of the alloy and the microrod still possesses amorphous structure without any ordering. The diffraction pattern for this sample is a halo ring shape which is typical for amorphous alloys. By increasing the forming temperature to near crystallisation temperature (496 ˚C), short range ordering in nanometer scales can be observed in some regions in the structure (Fig. 6.15a). However, the diffraction pattern verifies the amorphous structure after forming. The local ordering of atoms returns to the thermodynamic state of amorphous alloys. The amorphous structure is a thermodynamically unstable phase and atoms tend to become crystallised when the conditions is satisfied. By increasing temperature, the required time for ordering of atoms would significantly decreases as a result of higher diffusion coefficient and because of this local medium range ordering happens. However, due to short forming time, atoms are still unable to create long range ordering and become crystallised.

In addition, it is interesting to note that while all the diffraction patterns have ring shape, the diameter sizes vary with temperature. Figs 6-13b-6-15b display the diameter sizes of the ring in the diffraction pattern for the as-received and microrods at 450 ˚C and 496 ˚C. The ring diameter increases from 8.00 nm-1 before TPF to 8.46 nm-1 and 8.60 nm-1 after TPF at 450 ˚C and 496 ˚C, respectively. The differences between the as-received sample and microrods at 450 ˚C and 496 ˚C are 7.66% and

135

Micro Forming of MG Using a Microholed Die

9%, approximately. However, the difference between the diameters of the thermoplastically formed samples is near 1.3%. According to diffraction theory, the interatomic spacing is inversely corresponded to the halo ring diameter of the electron diffraction patterns [31, 32]. This demonstrates the reduction of interatomic distance after TPF. The lower interatomic space means that the free volume among atoms becomes less and material would be denser. The effect of this phenomenon on the mechanical properties will be discussed in the next Chapter.

Figure 6 .13 HRTEM analysis of the as-received MG

136

Micro Forming of MG Using a Microholed Die

Figure 6 .14 HRTEM analysis of the microrod fabricated at 450˚C

Figure 6 .15 HRTEM analysis of the microrod fabricated at 496 ˚C

6.4 Summary

In this Chapter a 500 µm microholed WC-Co die was fabricated by a diamond drilling tool and nanomachining. A number of experiments at different temperatures within SCLR are conducted to find out the effect of TPF on structure, apparent viscosity and mechanical properties. The force and time were 200 N and 60 s,

137

Micro Forming of MG Using a Microholed Die

respectively. Different microrods with different lengths were fabricated. With raising temperatures, the apparent viscosity of LM106a decreased significantly and reached to a minimum of 1.6 × 107 Pa.s at 496 ˚C. It was found that MG could not flow in to the die at temperatures lower than 450 ˚C and therefore this temperature would be the minimum applicable temperature for TPF of LM106a for a 500 µm die under 10

MPa. Highest molding temperature of around 496 ˚C is proposed as crystallisation is inhibited. Molding speed in the range of 10-3 is also proposed for TPF. The geometrical accuracy was very good for microrod and microribs and no difference was identified between the die and product geometry. The effects of oxidation pressure and capillary pressure were quantitatively calculated. It was shown that these factors have trivial effects on TPF of a 500 µm die. It was shown that the structure of microrods remained amorphous after TPF at all temperatures. However short range ordering could be observed in some regions of the sample thermoplastically formed at 496 ˚C. It was also found the diameter of the diffraction pattern in HRTEM increased after TPF. The diameter size becomes even larger with forming temperature rise. According to diffraction theory and considering the inverse relationship of diffraction pattern and interatomic spacing, it was concluded that interatomic spacing becomes smaller after TPF. This indicates free volume annihilation in the structure of MGs. Mechanical property of microrods and the effect of free volume annihilation on mechanical properties will be discussed in the next Chapter.

138

Mechanical Property Characterisation of MG after TPF

Chapter 7 Mechanical Property Characterisation of MG after TPF

139

Mechanical Property Characterisation of MG after TPF

TPF has been recognized as a promising manufacturing method for MG components, because it is based on the unique softening behaviour of MGs in their SCLR to enable the production of complex geometries at micro and nanoscales. The product quality by TPF of MG strongly depends on the forming temperature, stress, speed and dies. A higher forming temperature/stress or a smaller forming speed can enhance the formability of MGs in their SCLR, but at the greater risk of crystallisation. It is noted that the microstructure and properties of MGs are very sensitive to temperature and stress applied. For example, after annealing MGs became harder as a result of free volume annihilation or nucleation of nanocrystals

[159]. This annihilation can be attributed to structural relaxation [160] and/or crystallisation [44, 58, 101]. On the other hand shear deformation at high stresses loosens the atomic packing via the free volume creation [161-163], decreases viscosity and makes MG softer [101, 108]. Accordingly, in TPF where stress is applied at the SCLR, the property changes of MGs are likely to take place. However, so far no work has been done in order to characterise the mechanical property changes of MGs after TPF. Considering that the performance of a material in application is directly affected by their properties, improving or preserving the properties of MGs during TPF is essential.

Accordingly in this Chapter, the mechanical property changes of LM106a after TPF will be discussed. In Chapters 4, 5 and 6, TPF of LM106a by using different dies

(flat, microchannelled and microholed) and under different temperatures were conducted. In order to reveal the effect of TPF on MG property, in this Chapter mechanical properties of the products are examined by using microhardness and nanoindentation. According to the mechanical property results, viscosity variation, 140

Mechanical Property Characterisation of MG after TPF

TEM analyses, DSC results and free volume model, the mechanism of mechanical property enhancement after TPF is then identified.

7.1 Experimental procedure

After implementing TPF processes in Chapters 4-6, mechanical properties of the thermoplastically formed products at different temperatures and by using various dies were examined by using microhardness and nanoindentation instruments. For comparison, mechanical properties of the as-received MG are also investigated. Figs.

7.1 and 7.2 show the Struers Vickers microhardness tester (DuraScan-80 G5) and nanoindentaion facility (Hysiron TI-950 TriboIndenter) utilized for mechanical property measurements respectively. The microhardness tests were carried out using a loading time of 20s and a wide range of indentation loads from 1 Kgf to 10 kgf.

Before indentation all the samples were polished in order to remove any artifact due to the lack of flatness influencing the results. Indentation results are the average of five indentation tests under the same condition. After each test, the indentation mark is analysed using optical microscope equipped with the microhardness tester.

Nanoindentation is conducted on the sample under the constant load mode. The maximum load applied for nanoindentation was 10 mN. In order to find out the loading rate effect on the mechanical properties, different loading rates, 2×10-1 1/s,

5×10-2 1/s, 2×10-2 1/s and 5×10-3 1/s were utilized in this thesis. After nanoindentation tests, laod-displacement curves of the samples were analysed and compared with the one for the as-received material.

141

Mechanical Property Characterisation of MG after TPF

Figure 7 .1 Struers Vickers Microhardness tester (Dura scan-80 G5)

Figure 7 .2 Hysitron nanoindentation instrument

7.2 Microhardness results

In this section microhardness results of the samples thermoplastically formed on flat, microchannelled and microholed dies at different temperatures are revealed. For comparison microhardness results of the as-received material is also reported.

142

Mechanical Property Characterisation of MG after TPF

7.2.1 Flat die

Figure 7.3 illustrates the hardness result of the samples thermoplastically formed at

450˚C, 460˚C, 470˚C, 480˚C, 490˚C and 496˚C. The load and loading rates were

1Kgf and 5×10-2 1/s, respectively. For comparison the hardness of the as-received

MG is also shown as a line inside the diagram. It is evident that the hardness of the all thermoplastically formed samples is higher than the as received material; yet the hardness increases with temperature rise. The hardness increases from 464 HV for the as received materials to 498 HV for the sample thermoplastically formed at

496˚C. The error bars show the range of measurement variation for each sample. The trivial variations of error bars at each temperature confirm the validity of the observed hardness increase with temperature.

510

500

490

480

470

Hardness (HV1) Hardness 460 As –received material

450

440 440 450 460 470 480 490 500 Temperature (˚C)

Figure 7 .3 Hardness of the as-received materials and samples thermoplastically formed at different temperatures on flat die

143

Mechanical Property Characterisation of MG after TPF

Figures 7.4 a and b show typical indentation marks on the as as-received material at different loads (1kgf and 5 Kgf), in which one can clearly find the localised semicircular deformation bands near the indentation marks. These bands are parallel to each other. On the thermo-plastically formed samples, however, the localised deformation band no longer appears on the surface. As an example, the indentation mark on the sample formed at 450˚C (load = 1kgf) is shown in Fig. 7.5a.

The localised shear deformation observed near the indentation marks of the as- received MG are in fact slip step as a result of propagation of shear bands to the surface [164]. Shear bands take place when the applied pressure precedes the yield strength of the MG [90]. Considering the linear relation of hardness and yield strength [165, 166], it is understood that the materials with lower hardness would be more susceptible for shear band evolution. Due to direct relationship of hardness and yield stress, when the MG becomes harder yield stress also increases. Shear bands indicate onset of plasticity. Accordingly after hardening and increasing yield stress the shear bands will be disappeared. Thus the existence of shear bands in the as- received MG and lack of these bands on the surface of the thermoplastically formed samples confirm the higher hardness after TPF.

Semicircular shape of shear bands in MGs has been reported in literature [90, 164,

167, 168]. The shape of shear bands near indenter follow the contour of effective stress [44, 90]. Vaidyanathan et al. [90] showed that shear band traces on the surface of a Zr-based MG followed the Mohr-Coulomb criteria for yield stress. According to

144

Mechanical Property Characterisation of MG after TPF

their finite element model accompanied by experimental results, they established a model which was able to predict the semicircular shape of the shear bands.

Figure 7 .4 Optical microscope image of the indentation mark at a) load = 1 kgf and b) load = 5 kgf on the as-received material

145

Mechanical Property Characterisation of MG after TPF

Figure 7 .5 Optical microscope images of the indentation marks under (a) 1kgf, (b) 2 kgf (c) 5 kgf and (d) 10 kgf on the sample formed at 450˚C

It was reported that MG could be embrittled after heat treatment [169], which can be characterised by the evolving of crack near the edge of the indentation mark under a high indentation load [166]. To examine the impact of hardness improvement on brittleness, indentation tests are carried out with higher loads (2 Kgf, 5 Kgf and 10

Kgf) on the sample formed at 450˚C. As shown in Figs. 7.5b to 7.5d, no crack emerged around the indentation mark, indicating that the MG samples after TPF were not embrittled. Improving hardness can be accompanied by brittleness of MGs.

However in this thesis this topic was not deeply analysed, because the samples were small and conducting impact test on small samples were extremely difficult and sometimes not possible. However a simple brittleness analysis was conducted by

146

Mechanical Property Characterisation of MG after TPF

using microhardness test. Brittleness of MGs can be accompanied by emerging crack near indentation mark at high indentation loads. Accordingly, higher indentation loads (up to ten times higher) was applied during microhardness test to check the evolution of crack near indentation edge. However microscopic analysis of the thermoplastically formed sample at 450˚C (Fig 7.5) at higher loads did not show any crack on the surface. But systematic experiments on the brittleness of MGs after TPF are required to gain solid conclusion.

7.2.2 Microchannelled die

Figure 7.6 shows the hardness results of the microribs fabricated by TPF of MG on the microchannelled die at different temperatures. It is revealed that the hardness of the microribs is higher than the as-received material and increases from 477 HV at

450˚C to 497 HV at 496˚C. The hardness results of the microribs are very close to those of the flat samples indicating the trivial effect of die on the hardness of MGs after TPF.

147

Mechanical Property Characterisation of MG after TPF

510

500

490

480

470

Hardness (HV1) Hardness 460

450

440 440 450 460 470 480 490 500 Temperature (˚C)

Figure 7 .6 Hardness of the microribs fabricated at different temperatures

7.2.3 Microholed die

The hardness results of the micro rods fabricated via TPF of MGs at different temperatures are dedicated in Fig. 7.7. It is evident that the hardness of microrods increases by forming temperature rise. The hardness increases from 474 HV at

450˚C to 498 HV at 496˚C. The hardness result at each temperature is very close to those obtained for the other dies. Accordingly it is concluded that dies, at least in micrometer size, have trivial effect on mechanical property changes of MGs during

TPF.

148

Mechanical Property Characterisation of MG after TPF

510

500

490

480

470

Hardness (HV1) Hardness 460

450

440 440 450 460 470 480 490 500 Temperature (˚C)

Figure 7 .7 Hardness of the fabricated microrods at different temperatures

Considering the hardness results of the three dies, it is understood that the effect of die geometries on the hardness value is trivial. It is confirmed that fluid behaviour is independent of die size in micrometer scale in thermoplastic forming under constant temperature, pressure and loading time [60]. However, in nanometer scale due to the impact of different parameters such as capillary force fluid behaviour significantly changes with die size [72]. Meanwhile, according to the results of Figs.7.3, 7.6-7.7, it is found that the hardness of MG is independent of the die size in micrometer range (100micron>). Trivial effect of die geometry on fluid behaviour might be the primary reason of the hardness results in Figs. 7.3, 7.6-7.7.

149

Mechanical Property Characterisation of MG after TPF

7.3 Nanoindentation results

In this section nanoindentation results of the as-received materials and the MGs thermoplastically formed at different temperatures are presented. The effect of TPF on the load-displacement curve is discussed in section 7.3.1 and the loading rate effects on the load-displacement curve are explored in section 7.3.2.

7.3.1 The effect of TPF on load-displacement

Figures 7.8a-d illustrate the load-displacement curve of the as-received MG and thermoplastically formed samples at 460˚C, 480˚C and 496˚C under the loading rate of 2×10-2 1/s, respectively. The load displacement curve shows some serrations called pop-ins. The pop-ins is believed to be the effect of shear instability and localisation in the structure of MG during indentation.

150

Mechanical Property Characterisation of MG after TPF

Figure 7 .8 Load displacement curves of the a) as received and thermoplastically formed samples at b) 460˚C c) 480˚C d) 496˚C under the loading rate of 2×10-2 1/s

It is clear that the number of pop-ins in the load-displacement curve of the thermoplastically formed samples is lower than that of the as-received MG.

Furthermore, with increasing the forming temperature the pop-ins will be lowered in the charts. Since this pop-ins indicates shear localisation and instability, it is concluded that the sample thermoplastically formed would experience less instability during indentation. The reason behind this phenomenon will be discussed thoroughly in section 7.4.

151

Mechanical Property Characterisation of MG after TPF

7.3.2 Loading rate effect on load-displacement curve

In order to investigate the loading rate effect on the load displacement curves of the as-received as well as the thermoplastically formed samples, nanoindetnation tests were carried out at different loading rates, 2 ×10-1 1/s, 5 ×10-2 1/s, 2 ×10-2 1/s and 5

×10-3 1/s for each sample.

7.3.2.1 As-received MG

Figsures 7.9a-d show the load-displacement curve of LM106a at different loading rates. It is obvious that the pop-ins increase with loading rate reduction.

-1 Figure 7 .9 Load displacement curve of LM106a at loading rate of a) 2×10 1/s b) 5×10-2 1/s c) 2×10-2 1/s d) 5×10-3 1/s

152

Mechanical Property Characterisation of MG after TPF

7.3.2.2 Thermoplastically formed MG at 460˚C

Figures 7.10a-d show the load-displacement diagrams of the sample thermoplastically formed at 460˚C. Similar to as-received MG, the diagrams become smoother by increasing the loading rate, indicating the evolution of less pop-ins during indentation.

Figure 7 .10 Load displacement curve of the sample thermoplastically formed at 460˚C and loading rate of a) of 2×10-1 1/s b) 5×10-2 1/s c) 2×10-2 1/s d) 5×10-3 1/s

7.3.2.3 Thermoplastically formed sample at 480 ˚C

Nanoindentation analyses of the sample thermoplastically formed at 480˚C at different loading rates are shown in Figs. 7.11a-d. The trend of pop-in density with loading time is the same as the as-received MG and thermoplastically formed sample

153

Mechanical Property Characterisation of MG after TPF

at 460˚C. With increasing loading rate the number of serrations in the load displacement curve would be lowered. It is obvious that decreasing loading rate facilitates shear localisation in MGs.

Figure 7 .11 Load displacement curve of the sample thermoplastically formed at 480˚C at loading rate of a) 2×10-1 1/s b) 5×10-2 1/s c) 2×10-2 1/s d) 5×10-3 1/s

7.3.2.4 Thermoplastically formed sample at 496˚C

Figures 7.12a-d reveal the load displacement curves of the sample thermoplastically formed at 496˚C. This temperature is very close to the crystallization temperature

(497˚C) of LM106a. The diagrams are much smoother as compared with those of as- received MG and formed samples at lower temperatures. However, with decreasing

154

Mechanical Property Characterisation of MG after TPF

loading rate, shear localisation happens more in the curve. The detailed mechanism of this phenomenon will be discussed in section 7.4.

Figure 7 .12 Load displacement curve of the sample thermoplastically formed at 496˚C at loading rate of a) 2 × 10-1 1/s b) 5 × 10-2 1/s c) 2 × 10-2 1/s d) 5 × 10-3 1/s

7.4 Mechanism of mechanical property improvement

To reveal the mechanism of the hardness increase of the MG after TPF, the viscosity evolutions of the material during forming process were analysed. According to fluid mechanics, liquid with Reynolds number lower than 2,300 has a laminar behaviour and can be considered as Newtonian fluid [151], in which the Reynolds number Re is

3 defined by ρvDe/η, where ρ is the liquid density (6.7 g/cm for LM106a), v is the forming velocity (here in µm/s range), De is the sample diameter (here is in millimeter range), and η is the liquid viscosity (107 Pa.s ~ 109 Pa.s for LM106a in 155

Mechanical Property Characterisation of MG after TPF

this study) [150]. It is obvious that Reynolds number is <<1. Hence, the viscosity of the MG behaviour in TPF process can be calculated by = /3 [120, 151],

𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 where σ is stress and is strain rate. 𝜂𝜂 𝜎𝜎 𝜀𝜀̇

𝜀𝜀̇

Figure 7.13 presents the viscosity changes of the MG during the forming processes at different temperatures. At lower temperatures (i.e. 450˚C, 460˚C, 470˚C and

480˚C), where the viscosity is far from the equilibrium state, the viscosity is continually increased with time. At higher temperatures (496˚C and 490˚C), however, the viscosity reaches a plateau during the test. This indicates that the viscosity is approaching to its equilibrium state (ηeq). In addition, the viscosity variation shows less serration amplitude in the diagrams at higher temperatures, particularly at 496˚C.

Spaepen [101, 152] reported the similar increase of viscosity with time in compressing a Pd-based MG within its SCLR. It is found that the slope of the viscosity-time chart changes from linear at glassy state to nonlinear in SCLR [152].

Near the equilibrium state the viscosity reaches a plateau at which the change rate becomes zero [100, 152]. This trend was attributed to the variation of free volume evolution in the material.

156

Mechanical Property Characterisation of MG after TPF

20000

2000

200 Log (Viscosity (MPa.s)) Log (Viscosity 450 ˚C 480 ˚C

460 ˚C 490 ˚C

470 ˚C 496 ˚C

20 0 10 20 30 40 50 60 time (s)

Figure 7.13 Logarithmic variation of viscosity of LM106a with time at different forming temperatures

Free volume in MGs plays a key role in their behaviour and properties [102, 169,

170]. TPF may manipulate the free volume concentration depending on the working temperature, strain, stress and strain rate [100, 101, 171]. Free volume in MGs can be described by

= exp (7-1) γυ∗ 𝑐𝑐𝑓𝑓 � υf � where cf is the free volume concentration, γ is a geometrical value ranging between

* 0.5 and 1, υf is the average free volume per atom and υ is the critical size for atomic jump [105]. At a given temperature, it was found that viscosity and free volume

157

Mechanical Property Characterisation of MG after TPF

concentration are inversely corresponded, i.e., [101]. Thus according to 1 𝜂𝜂 ∝ − 𝑐𝑐𝑓𝑓 Fig. 7.31, it is obvious that the free volume concentration is decreasing during TPF.

The change of free volume concentration during TPF could be attributed to structural relaxation [100, 101, 105]. In SCLR, the non-equilibrium microstructure of the MG tends to relax towards an equilibrium microstructure, and accordingly the free volume concentration decreases [44, 100, 101, 171]. The rate of free volume annihilation as a result of the structural relaxation can be described [101]:

( , ) 7-2) 𝑑𝑑𝑐𝑐𝑓𝑓 𝑑𝑑𝑑𝑑 ∝ 𝑐𝑐𝑓𝑓 − 𝑐𝑐𝑓𝑓 𝑒𝑒𝑒𝑒 where cf,eq is the equilibrium free volume concentration at each temperature.

Accordingly, the change rate of viscosity [152]:

(7-3) 𝑑𝑑𝑑𝑑 𝑑𝑑𝑑𝑑 ∝ �𝜂𝜂𝑒𝑒𝑒𝑒 − 𝜂𝜂� According to Eq. (7-3), when , d /d 0, and the viscosity increases with

𝑒𝑒𝑒𝑒 time as shown in Fig. 7.13 at 𝜂𝜂low ≫temperatures.𝜂𝜂 𝜂𝜂 𝑡𝑡 ≫ When approaches , however,

𝑒𝑒𝑒𝑒 d /d 0. At a high temperature, the structural relaxation𝜂𝜂 process is very𝜂𝜂 fast [100,

101,𝜂𝜂 105]𝑡𝑡 → , and thus the free volume concentration can quickly reach its equilibrium state. In this case, the viscosity can approach its equilibrium value very fast, as shown in Fig. 7.13.

DSC analysis results of the as-received and thermoplastically formed samples at

450˚C and 496˚C are shown in Fig. 7.14. The shaded area is the heat release as a result of structural relaxation during DSC tests. This is the standard method of

158

Mechanical Property Characterisation of MG after TPF

determining heat release as a result of structural relaxation before (Tg) for MGs. The start of structural relaxation is the point that gradient start changing for as-received

MG (around 200˚C beginning of the shade area). The end of relaxation is the beginning of crystallisation (end of shaded area). The area below the graph (shaded area) is considered as the heat release as a result of structural relaxation. The formulation is based on total enthalpy change (ΔH) during DSC test. The exact formulation detail can be found in Ref [153] of the thesis. For thermoplastically formed samples less heat is released during the DSC test. In addition, relaxation enthalpy changes between the thermoplastically formed and the as-received MGs increase with temperature rise. Since relaxation enthalpy change and free volume reduction directly correlate to each other [153], free volume annihilation during TPF is confirmed.

Figure 7.14 DSC results of the as-received and thermoplastically formed MGs at 450˚C and 496˚C

159

Mechanical Property Characterisation of MG after TPF

It is clear that, during the structural relaxation free volume concentration would decrease, and thus the viscosity would increase, matching the results at lower temperatures in Fig. 7.13. At higher temperatures, the structural relaxation is very fast, and thus the free volume concentration can quickly reach its equilibrium state.

In this case, the viscosity variation will be very small, matching the results at higher temperatures in Fig. 7.13.

It is noted that visco-plastic deformation during the forming process may create the free volume [100, 101]. However, as shown in Eq. (7-1), the free volume can be created only if the interatomic distance is larger than the critical size for atomic jump

υ*. Accordingly, at low stresses where the interatomic space would not reach the critical value, free volume would not be created [100, 101, 106]. It was reported that for a Pd-based MG under the pressure of 15 MPa at 555 K, the free volume creation is negligible, and its viscosity increases with time [100, 101]. Considering that the forming pressures applied in our study are very low (< 10 MPa), the contribution of plastic deformation on the changes of free volume is negligible.

It is noted that the free volume concentration is directly related to the interatomic spacing, which can be characterized by the diameter of the innermost ring of the observed halo pattern in the electron diffraction pattern [31]. The increase of the diameter after TPF in Figs. 4.21- 4.23 confirm that the interatomic spacing and accordingly free volume concentration decrease [32, 172], leading to the hardness increase of MGs.

160

Mechanical Property Characterisation of MG after TPF

Nanoindentation result analyses of MGs would also confirm the free volume annihilation after TPF. It is believed that serrations in the load displacement curves of MGs indicate shear localization. Thus shear localization would decrease after

TPF. It means that TPF improves the homogeneity of the MG. Shear localization occurs from the weaker sites of the alloy. Free volumes are among the weakest zones of an alloy with lower strength. It means that the MG with higher free volume concentration is more susceptible for shear localization. As mentioned earlier during the TPF, the free volume relaxation would decrease as a result of structural relaxation. Therefore, free volume annihilation reduces the preferred site of shear localization and serrations in load displacement curves. With increasing the TPF temperature due to faster structural relaxation, as previously proved by viscosity variation and TEM analysis, the interatomic spacing would be reduced. The lower interatomic spacing improves the strength of the MG in front of shear localization.

Thus, with increasing the TPF temperature the number of serration would be more decreased.

Loading rate dependence of MGs has also been corresponded to shear localization.

Schuh et al. [43, 44] investigated the plastic deformation and serrated flow of MGs.

Based on this, serrations are more prominent at lower loading rate and with increasing loading rates, these serration become less evident or even disappear in the curve [43].

The nature of serrations in nanoindentation would be the result of shear localisation

[43]. Onset of plasticity of MGs in nanoindentation occurs in pop-ins [43]. The reason behind the appearance of pop-ins in the curves with decreasing loading rate is 161

Mechanical Property Characterisation of MG after TPF

because of changing the mechanism of plastic deformation of MGs. At low loading rate discrete yielding (multiple shear bands) occurs during plastic deformation, while at high loading rates continuous yielding (single shear band) happens and no deformation burst is observed. Deformation rate affects shear band formation in

MGs which results in the various serration types in different MGs.

In nanoindentation, it is believed that at high loading rates, a single shear band carries the plastic deformation, but at low loading rate, multiple shear bands simultaneously operate and lead to plastic deformation [43, 173]. During indentation, the first response of MGs would be elastic deformation and then at a critical stress level, yielding happens [43]. At this stage the first shear band is activated in the structure and if the loading rate is low, this shear band would have sufficient time to accommodate the applied strain and results in the strain burst. On the other hand, if the loading rate is higher than the relaxation rate of a single shear band, the shear band will be operating and when the stress level exceeds the yield criterion, the next shear band will be activated in the structure. At low loading rate, multiple shear bands are activated in order to accommodate the plastic strain [43].

Statistical analyses of load displacement curves were conducted. Statistical analyses of load displacement curves were conducted. First, shear bands were considered as a location with zero gradients to quantify them. However, the shear bands could not be quantified exactly by this method. In fact, the shear bands in load displacement graphs have gradients close to zero, but not zero exactly. The analysis was then preceded on different small areas near obvious shear bands. However, it was found that the slope and length of shear bands were not unique and changed throughout the 162

Mechanical Property Characterisation of MG after TPF

experiments and no solid conclusion could be obtained, accordingly. This can be an interesting topic for future research in this area.

The Young’s modulus and hardness increased with molding temperatures. Young’s modulus increased from 96 GPa for the as-received MG to around 106 GPa after

TPF. Hardness also increased from 5.8 GPa to around 7 GPa after TPF. This also confirms hardening and free volume annihilation of MGs after TPF.

7.5 Summary

In this Chapter the mechanical properties of LM106a after TPF is thoroughly investigated via microhardness and nanoindentation techniques. It is found that the

MGs become harder after TPF and the hardening increases with temperature rise.

Nanoindentation analyses also reveal the less serration in load-displacement curve of

LM106a after TPF; yet the number of serrations is reduced with increasing forming temperature. It is found that serrations are quite sensitive to loading rate. With increasing loading rate, the curve becomes smoother with lower serrations.

According to viscosity variation, micro hardness and nanoindentation results, DSC examinations and TEM analyses at different temperatures, free volume annihilation during TPF of MG is verified. Structural relaxation is considered as the mechanism of free volume annihilation and mechanical property enhancement during forming.

The lower density of free volume after forming results in the hardening of MG and lower pop-ins in the load displacement curve. Loading rate dependency of MG

163

Mechanical Property Characterisation of MG after TPF

before and after TPF is verified. The primary reason behind the loading rate dependency of MG is the activation of shear localization in different zones, depending on the loading rate. With increasing loading rate, shear localization is appeared in one site of the MG which causes less serration in load-displacement curve. However, by decreasing the loading rate, shear localization is activated in different sites of the MGs leading to more serrations in load displacement curve.

164

Conclusions and Future Research

Chapter 8 Conclusions and Future Research

165

Conclusions and Future Research

8.1 Conclusions

This thesis conducted an in-depth investigation in TPF of MGs. The following techniques were used to achieve the aims of the thesis: glass moulding machining, nanomachining, optical microscopy, FIB, SEM/HRSEM, HRTEM, EDS, DSC,

XRD, Zygo roughness analysis, microhardness and nanoindentation tests.

The primary conclusions of this thesis are as follow:

1) PTFE, WC-Co and sapphire were found to be the materials with lowest

adhesion with MGs. However PTFE is not appropriate for all MGs due to

low melting point. WC-Co has better machinability and can provide deeper

feature than can sapphire. Thus, WC-Co was selected as the die materials.

SFE and BDE were identified as the key factors indicating the adhesion

status. A new approach was developed to calculate the SFE of MGs. The SFE

results were verified by the data in literature.

2) It was found that chemical reaction occurs at the MG/die interface and this

was verified by HRSEM. Diffusion of element (from MG to die and vice

versa) was also revealed and was verified via EDS elemental analyses.

Chemical reaction and diffusion were identified as the primary mechanisms

of MG/die adhesion.

3) A novel model was developed to calculate MG/die work of adhesion in TPF.

The model was verified by the experimental results of this thesis.

4) Microribs and microrods were fabricated successfully via TPF of MGs using

microchannelled and microholed WC-Co dies at different temperatures. No

166

Conclusions and Future Research

adhesion occurred between the microribs and microrods with dies and they

could be easily detached from the dies.

5) The microforming of the MGs using microholed die at different temperatures

demonstrated that formability of LM106a could be increased until near

crystallisation temperature. The minimum temperature allowing the

supercooled metal flow into the die was 450˚C. The apparent viscosity varied

from 6.6 × 1010 at 450 ˚C Pa.s to 1.6 × 107 Pa.s at 496˚C. It was also

demonstrated that the effects of capillary pressure and oxidation on apparent

viscosity are trivial in microforming of MG and can be ignored.

6) Structural analyses of MGs after TPF by using XRD and HRTEM verified

the amorphous structure of MGs after TPF. However at 496˚C short range

ordering was identified in some regions of LM106a. It was revealed that the

diffraction pattern of MGs changes after TPF. The diameter of the first ring

of diffraction pattern increases after TPF; yet the diameter increases with

temperature increases. Considering the relationship of the ring diameter with

the interatomic spacing in diffraction theory, it was confirmed that

interatomic spacing reduces after TPF which means free volume annihilation

occurs during this process.

7) The DSC results demonstrate that the heat release of the samples

thermoplastically formed were smaller than the as-received MG. This

confirms the occurrence of structural relaxation during TPF. Considering the

relationship between relaxation enthalpy and free volume, it is understood

that free volume annihilation occur during MG forming.

8) Viscosity variation during TPF and diffraction pattern analyses confirmed

free volume annihilation. At lower temperatures (i.e. 450 ˚C, 460 ˚C, 470 ˚C 167

Conclusions and Future Research

and 480 ˚C) where apparent viscosity is far from equilibrium, its magnitude

increased during the tests. However, at higher temperatures where the system

approaches its equilibrium state, viscosity reached a plateau and remained

approximately constant.

9) Mechanical property assessments of the thermoplastically formed MGs on

different dies i.e. flat, microchannelled and microholed, at the same

temperature and pressure indicate the trivial effect of die features on

mechanical properties of MGs after TPF.

10) It was found that the hardness of MGs increases after TPF. The hardening

effect enhances with forming temperature. The hardness of LM106a

increased from 464 HV in as-received state to 498 HV after TPF at 496˚C. In

addition, the circular shear localisation bands around the indentation mark of

the as-received MG were revealed. However, after TPF, the shear bands

disappeared on the surface of MG.

11) Nanoindentation analyses showed less pop-ins in load-displacement curves

after TPF. The number of pop-ins decreased with increasing forming

temperature. The direct relationship between pop-ins and free volume

concentration indicates the lower free volume concentration after TPF.

12) The number of pop-ins increased with decreasing loading rates. At low

loading rates multiple pop-ins were observed in load-displacement diagram.

But at higher loading rates smooth graphs achieved in load displacement

graphs.

13) Considering the viscosity variation, HRTEM analyses, DSC results, and the

microhardness and nanoindentation analyses, it was demonstrated that free

168

Conclusions and Future Research

volume annihilation through structural relaxation is the cause of the

hardening of MGs after TPF.

8.2 Future research

This thesis conducted an in-depth investigation of TPF of MGs. However increasing attention of these alloys in science and technology necessitates further investigations.

The following topics are suggested as potential research areas.

 The size dependence of mechanical properties such as stress-strain curve and

viscosity for MGs has been demonstrated in the literatures. However, the

effect of die size in nanometric scale on the mechanical properties of MGs

after TPF has not yet been investigated. Due to increasing importance of

nanotechnology, it would be necessary to examine the size dependence of

mechanical properties (e.g. hardness) in nanometric scales.

 Although in this thesis mechanical and structural properties of MGs were

investigated, due to applications of these alloys in different fields it would be

necessary to investigate physical, chemical, magnetic, biocompatibility and

optical properties of MGs after TPF. In this way deeper understanding of the

effect of TPF on different materials properties would be achieved.

 MGs are basically brittle alloys with low fracture toughness compared with

metals. In this thesis it was shown that the hardness of the MG increased after

TPF. Hardening sometimes leads to embrittlement of MG. It would be

interesting to conduct fracture toughness tests on the MGs before and after

TPF in order to examine the effect of hardening on the fracture toughness of

the alloys.

169

Conclusions and Future Research

 Molecular dynamic simulation is a power instrument for characterising the

materials behaviour in nano scales. Experimentally the mechanical property

enhancement was demonstrated in this thesis. However, molecular dynamic

can be utilized to give a deeper view regarding atomic position, shear

localisation, structural relaxation, free volume annihilation, density change

and hardening during TPF.

 Friction coefficient would be one of the external factors which might affect

the apparent viscosity. Conducting wear test at different temperatures within

SCLR to establish the friction coefficient changes can be used as a potential

topic for future research.

170

Conclusions and Future Research

List of Publications

Journal papers

1- Amir Monfared, Weidong Liu and Liangchi Zhang, “Metallic glass hardening after thermoplastic forming”, Materials Science and Engineering A 725 (2018) 181-

186.

2- Amir Monfared, Weidong Liu and Liangchi Zhang, “On the adhesion between metallic glass and dies during thermoplastic forming”, Journal of Alloys and

Compounds 711 (2017) 235-242

3- Amir Monfared, Weidong Liu and Liangchi Zhang, “Thermomechanical adhesion between metallic glass and die materials”, Materials Science Forum 879 (2017)

1323-1327

4- Amir Monfared, Weidong Liu and Liangchi Zhang, “Microforming of a Zr-based metallic glass at different temperatures: Apparent viscosity, structure and mechanical properties”, under preparation

5- Amir Monfared, Weidong Liu and Liangchi Zhang, “Nanoindentation analyses of the thermoplastically formed metallic glass at different temperatures”, under preparation

Conference

Amir Monfared, Weidong Liu and Liangchi Zhang, “Thermomechanical adhesion between metallic glass and die materials” presented in the international conference

171

Conclusions and Future Research

on processing and manufacturing of advanced materials, Processing, Fabrication,

Properties, Applications May 29- June 3 2016, Graz, Austria.

172

References

1. Suryanarayana, C. and A. Inoue, Bulk metallic glasses. 2011: CRC Press.

2. Wang, W.-H., C. Dong, and C. Shek, Bulk metallic glasses. Materials

Science and Engineering: R: Reports, 2004. 44(2): p. 45-89.

3. Liu, W. and L. Zhang, On the nano/micro-mechanics of metallic glass.

Critical Reviews in Solid State and Materials Sciences, 2015. 40(3): p. 137-

163.

4. Rahaman, M., L. Zhang, and H. Ruan, Effects of environmental temperature

and sliding speed on the tribological behaviour of a Ti-based metallic glass.

Intermetallics, 2014. 52: p. 36-48.

5. Ashby, M. and A. Greer, Metallic glasses as structural materials. Scripta

Materialia, 2006. 54(3): p. 321-326.

6. Kumar, G., A. Desai, and J. Schroers, Bulk metallic glass: the smaller the

better. Advanced materials, 2011. 23(4): p. 461-476.

7. Schroers, J., Processing of bulk metallic glass. Advanced materials, 2010.

22(14): p. 1566-1597.

8. Bakkal, M., et al., Machining of a Zr–Ti–Al–Cu–Ni metallic glass. Scripta

Materialia, 2004. 50(5): p. 583-588.

9. Jiang, M. and L. Dai, Formation mechanism of lamellar chips during

machining of bulk metallic glass. Acta Materialia, 2009. 57(9): p. 2730-2738.

10. Kumar, G., et al., Atomically smooth surfaces through thermoplastic forming

of metallic glass. Applied Physics Letters, 2010. 97(10): p. 101907.

173

References

11. Stern, E.M., Roman glassblowing in a cultural context (Commerce, society,

imperial Rome). American Journal of Archaeology, 1999. 103(3): p. 441-

484.

12. Klement, W., R.H. Willens, and P. Duwez, Non-Crystalline Structure in

Solidified Gold-Silicon Alloys. Nature, 1960. 187(4740): p. 869-870.

13. Inoue, A., Stabilization of metallic supercooled liquid and bulk amorphous

alloys. Acta Materialia, 2000. 48(1): p. 279-306.

14. Johnson, W.L., Bulk amorphous metal - An emerging engineering material.

Jom-Journal of the Minerals Metals & Materials Society, 2002. 54(3): p. 40-

43.

15. Wang, W.H., C. Dong, and C.H. Shek, Bulk metallic glasses. Materials

Science & Engineering R-Reports, 2004. 44(2-3): p. 45-89.

16. Davies, H.A., Amorphous Metallic Alloys, ed. F.E. Luborsky. 1983, London:

Butterworths.

17. Johnson, W.L., Thermodynamic and Kinetic Aspects of the Crystal to Glass

Transformation in Metallic Materials. Progress in Materials Science, 1986.

30(2): p. 81-134.

18. Luborsky, F.E., Amorphous Metallic Alloys, ed. F.E. Luborsky. 1983,

London: Butterworths.

19. Chen, H.S., Thermodynamic Considerations on Formation and Stability of

Metallic Glasses. Acta Metallurgica, 1974. 22(12): p. 1505-1511.

20. Turnbull, D. and J.C. Fisher, Rate of nucleation in condensed systems.

Journal of Chemical Physics, 1949. 17: p. 71-73.

21. Drehman, A.J., A.L. Greer, and D. Turnbull, Bulk formation of a metallic

glass: Pd40Ni40P20 Applied Physics Letters, 1982. 41: p. 716-717. 174

References

22. Kui, W.H., A.L. Greer, and D. Turnbull, Formation of Bulk Metallic Glass by

Fluxing. Applied Physics Letters, 1984. 45: p. 615-616.

23. Inoue, A., T. Zhang, and T. Masumoto, Al–La–Ni amorphous-alloys with a

wide supercooled liquid region. Materials Transactions Jim, 1989. 30(12): p.

965-972.

24. Inoue, A., et al., Mg-Cu-Y Bulk Amorphous Alloys with High Tensile Strength

Produced by a High-Pressure Die Casting Method. Materials Transactions

Jim, 1992. 33: p. 937-935.

25. Chen, M., A brief overview of bulk metallic glasses. NPG Asia Materials,

2011. 3: p. 82.

26. Sun, Y.T., et al., Machine Learning Approach for Prediction and

Understanding of Glass-Forming Ability. The journal of physical chemistry

letters, 2017. 8(14): p. 3434-3439.

27. Cavagna, A., Supercooled liquids for pedestrians. Physics Reports-Review

Section of Physics Letters, 2009. 476(4-6): p. 51-124.

28. Scherer, G.W., Theories of Relaxation. Journal of Non-Crystalline Solids,

1990. 123(1-3): p. 75-89.

29. Lan, S., et al., Hidden amorphous phase and reentrant supercooled liquid in

Pd-Ni-P metallic glasses. Nature Communications, 2017. 8: p. 14679.

30. Bates, R.L. and J.A. Jackson, Dictionary of Geological Terms: Third Edition

(Rocks, Minerals and Gemstones). 1984, USA: American Geological

Institute.

31. Bian, X., et al., Manipulation of free volumes in a metallic glass through Xe-

ion irradiation. Acta Materialia, 2016. 106: p. 66-77.

175

References

32. Deng, Y., et al., HRTEM analysis of nanocrystallization during uniaxial

compression of a bulk metallic glass at room temperature. Ultramicroscopy,

2004. 98(2): p. 201-208.

33. Liu, W.D., H.H. Ruan, and L.C. Zhang, Atomic rearrangements in metallic

glass: their nucleation and self-organization Acta Materialia, 2013. 61(16):

p. 6050-6060.

34. Claridge, S.A., et al., Cluster-assembled materials. ACS Nano, 2009. 3(2): p.

244-55.

35. Evertz, S., et al., Thermal expansion of Pd-based metallic glasses by ab initio

methods and high energy X-ray diffraction. Scientific Reports, 2017. 7(1): p.

15744.

36. Faupel, F., et al., Diffusion in metallic glasses and supercooled melts.

Reviews of modern physics, 2003. 75(1): p. 237.

37. Waseda, Y. and K. Aust, Corrosion behaviour of metallic glasses. Journal of

Materials Science, 1981. 16(9): p. 2337-2359.

38. Xu, T., S. Pang, and T. Zhang, Glass formation, corrosion behavior, and

mechanical properties of novel Cr-rich Cr–Fe–Mo–C–B–Y bulk metallic

glasses. Journal of Alloys and Compounds, 2015. 625: p. 318-322.

39. Li, H., et al., Biodegradable Mg–Zn–Ca–Sr bulk metallic glasses with

enhanced corrosion performance for biomedical applications. Materials &

Design, 2015. 67: p. 9-19.

40. Trexler, M.M. and N.N. Thadhani, Mechanical properties of bulk metallic

glasses. Progress in Materials Science, 2010. 55(8): p. 759-839.

176

References

41. Park, K.-W., et al., Deformation behaviors under tension and compression:

Atomic simulation of Cu 65 Zr 35 metallic glass. Intermetallics, 2011. 19(8):

p. 1168-1173.

42. Zhang, Q., Q.-K. Li, and M. Li, Key factors affecting mechanical behavior of

metallic glass nanowires. Scientific Reports, 2017. 7: p. 41365.

43. Schuh, C. and T. Nieh, A nanoindentation study of serrated flow in bulk

metallic glasses. Acta Materialia, 2003. 51(1): p. 87-99.

44. Schuh, C.A., T.C. Hufnagel, and U. Ramamurty, Mechanical behavior of

amorphous alloys. Acta Materialia, 2007. 55(12): p. 4067-4109.

45. Wang, G., et al., Fatigue behavior of bulk-metallic glasses. Intermetallics,

2004. 12(7): p. 885-892.

46. Malekan, M., et al., Formation of bulk metallic glass in situ nanocomposite

in (Cu 50 Zr 43 Al 7) 99 Si 1 alloy. Materials Science and Engineering: A,

2012. 553: p. 10-13.

47. Malekan, M., et al., Effect of Ge addition on mechanical properties and

fracture behavior of Cu–Zr–Al bulk metallic glass. Journal of Alloys and

Compounds, 2009. 484(1): p. 708-711.

48. Eckert, J., et al., Mechanical properties of bulk metallic glasses and

composites. Journal of materials research, 2007. 22(2): p. 285-301.

49. Johnson, W., Bulk amorphous metal—An emerging engineering material.

Jom, 2002. 54(3): p. 40-43.

50. Sekol, R.C., et al., Bulk metallic glass micro fuel cell. Small, 2013. 9(12): p.

2081-2085.

51. Elias, C., et al., Biomedical applications of titanium and its alloys. Jom,

2008. 60(3): p. 46-49. 177

References

52. Li, H. and Y. Zheng, Recent advances in bulk metallic glasses for biomedical

applications. Acta biomaterialia, 2016. 36: p. 1-20.

53. Schroers, J. and G. Kumar, Bulk Metallic Glass in Micro to Nano Length

Scale Applications. The Nano-Micro Interface: Bridging the Micro and Nano

Worlds, 2015: p. 159-188.

54. Hofmann, D.C. and S.N. Roberts, Microgravity metal processing: from

undercooled liquids to bulk metallic glasses. npj Microgravity, 2015. 1: p.

15003.

55. Ramasamy, P., et al., High pressure die casting of Fe-based metallic glass.

Scientific reports, 2016. 6: p. 35258.

56. Yeo, S.H., et al., Processing of zirconium-based bulk metallic glass (BMG)

using micro electrical discharge machining (Micro-EDM). Materials and

Manufacturing Processes, 2009. 24(12): p. 1242-1248.

57. Bakkal, M., A.J. Shih, and R.O. Scattergood, Chip formation, cutting forces,

and tool wear in turning of Zr-based bulk metallic glass. International

Journal of Machine Tools and Manufacture, 2004. 44(9): p. 915-925.

58. Li, N., W. Chen, and L. Liu, Thermoplastic micro-forming of bulk metallic

glasses: A review. JOM, 2016. 68(4): p. 1246-1261.

59. Li, N., et al., A thermoplastic forming map of a Zr-based bulk metallic glass.

Acta Materialia, 2013. 61(6): p. 1921-1931.

60. Chiu, H.M., et al., Thermoplastic extrusion of bulk metallic glass. Scripta

Materialia, 2009. 61(1): p. 28-31.

61. Ishida, M., et al., Wear resistivity of super-precision microgear made of Ni-

based metallic glass. Materials Science and Engineering: A, 2007. 449: p.

149-154. 178

References

62. Shao, Z., et al., Size-dependent viscosity in the super-cooled liquid state of a

bulk metallic glass. Applied Physics Letters, 2013. 102(22): p. 221901.

63. Saotome, Y. and A. Inoue. Superplastic micro-forming of microstructures. in

Micro Electro Mechanical Systems, 1994, MEMS'94, Proceedings, IEEE

Workshop on. 1994. IEEE.

64. Saotome, Y., et al., The micro-formability of Zr-based amorphous alloys in

the supercooled liquid state and their application to micro-dies. Journal of

Materials Processing Technology, 2001. 113(1): p. 64-69.

65. Bardt, J.A., et al., Micromolding three-dimensional amorphous metal

structures. Journal of materials research, 2007. 22(2): p. 339-343.

66. Schroers, J., On the formability of bulk metallic glass in its supercooled

liquid state. Acta Materialia, 2008. 56(3): p. 471-478.

67. Rieser, D., G. Spieß, and P. Manns, Investigations on glass-to-mold sticking

in the hot forming process. Journal of Non-Crystalline Solids, 2008. 354(12):

p. 1393-1397.

68. Sun, H., et al., Anti-sticking treatment for a nanoimprint stamp. Applied

Surface Science, 2008. 254(10): p. 2955-2959.

69. Beck, M., et al., Improving stamps for 10 nm level wafer scale nanoimprint

lithography. Microelectronic Engineering, 2002. 61: p. 441-448.

70. Navabpour, P., et al., Evaluation of non-stick properties of magnetron-

sputtered coatings for moulds used for the processing of polymers. Surface

and Coatings Technology, 2006. 201(6): p. 3802-3809.

71. Barbero, D.R., et al., High-resolution nanoimprinting with a robust and

reusable polymer mold. Advanced Functional Materials, 2007. 17(14): p.

2419. 179

References

72. Kumar, G., H.X. Tang, and J. Schroers, Nanomoulding with amorphous

metals. Nature, 2009. 457(7231): p. 868-872.

73. Schroers, J., Q. Pham, and A. Desai, Thermoplastic forming of bulk metallic

glass—a technology for MEMS and microstructure fabrication.

Microelectromechanical Systems, Journal of, 2007. 16(2): p. 240-247.

74. Saiz, E., R.M. Cannon, and A.P. Tomsia, High-temperature wetting and the

work of adhesion in metal/oxide systems. Annu. Rev. Mater. Res., 2008. 38:

p. 197-226.

75. von Fraunhofer, J.A., Adhesion and cohesion. International journal of

dentistry, 2012. 2012.

76. Suh, J.-Y., et al., Novel thermoplastic bonding using a bulk metallic glass

solder. Scripta Materialia, 2008. 59(8): p. 905-908.

77. Wang, J., et al., Diffusion bonding of a Zr-based metallic glass in its

supercooled liquid region. Intermetallics, 2014. 46: p. 236-242.

78. Kawamura, Y., Liquid phase and supercooled liquid phase welding of bulk

metallic glasses. Materials Science and Engineering: A, 2004. 375: p. 112-

119.

79. Quéré, D., Surface chemistry: Fakir droplets. Nature materials, 2002. 1(1): p.

14-15.

80. De Gennes, P.-G., Wetting: statics and dynamics. Reviews of modern

physics, 1985. 57(3): p. 827.

81. Bennett, I., J. Kranenburg, and W. Sloof, Modeling the influence of reactive

elements on the work of adhesion between oxides and metal alloys. Journal of

the American Ceramic Society, 2005. 88(8): p. 2209-2216.

180

References

82. Gerberich, W. and M. Cordill, Physics of adhesion. Reports on Progress in

Physics, 2006. 69(7): p. 2157.

83. Ding, S., J. Kong, and J. Schroers, Wetting of bulk metallic glass forming

liquids on metals and ceramics. Journal of Applied Physics, 2011. 110(4): p.

043508.

84. Zhang, Q., Q.-K. Li, and M. Li, Processing dependence of mechanical

properties of metallic glass nanowires. Applied Physics Letters, 2015.

106(7): p. 071905.

85. Callister, W.D. and D.G. Rethwisch, Materials science and engineering. Vol.

5. 2011: John Wiley & Sons NY.

86. Lu, J., G. Ravichandran, and W.L. Johnson, Deformation behavior of the Zr

41.2 Ti 13.8 Cu 12.5 Ni 10 Be 22.5 bulk metallic glass over a wide range of

strain-rates and temperatures. Acta materialia, 2003. 51(12): p. 3429-3443.

87. Qiao, J. and J. Pelletier, Dynamic mechanical relaxation in bulk metallic

glasses: a review. Journal of Materials Science & Technology, 2014. 30(6):

p. 523-545.

88. Liu, W.-D., H.-H. Ruan, and L.-C. Zhang, Understanding the brittleness of

metallic glasses through dynamic clusters. Journal of Materials Research,

2014. 29(4): p. 561-568.

89. Huang, R., et al., Inhomogeneous deformation in metallic glasses. Journal of

the Mechanics and Physics of Solids, 2002. 50(5): p. 1011-1027.

90. Vaidyanathan, R., et al., Study of mechanical deformation in bulk metallic

glass through instrumented indentation. Acta Materialia, 2001. 49(18): p.

3781-3789.

181

References

91. Golovin, Y.I., et al., Serrated plastic flow during nanoindentation of a bulk

metallic glass. Scripta materialia, 2001. 45(8): p. 947-952.

92. Huang, Y., et al., Bending behavior of TiZrNiCuBe bulk metallic glass.

Journal of Alloys and Compounds, 2012. 541: p. 359-364.

93. Chen, Y., et al., Thermoplastic deformation and micro/nano-replication of an

Au-based bulk metallic glass in the supercooled liquid region. Materials

Science and Engineering: A, 2012. 556: p. 488-493.

94. Lee, K.S., et al., Thermomechanical characterization of Cu 47.5 Zr 47.5 Al 5

bulk metallic glass within the homogeneous flow regime. Intermetallics,

2009. 17(1): p. 65-71.

95. Deng, D., et al., Creep and structural relaxation in Pd40Ni40P20 glass. Acta

metallurgica et materialia, 1993. 41(4): p. 1089-1107.

96. Schuh, C.A., T.C. Hufnagel, and U. Ramamurty, Overview No.144 -

Mechanical behavior of amorphous alloys. Acta Materialia, 2007. 55(12): p.

4067-4109.

97. Cohen, M.H. and D. Turnbull, Molecular transport in liquids and glasses.

Journal of Chemical Physics, 1959. 31: p. 1164-1169.

98. Polk, D.E. and D. Turnbull, Flow of Melt and Glass Forms of Metallic

Alloys. Acta Metallurgica, 1972. 20: p. 493-498.

99. Spaepen, F., A microscopic mechanism for steady state inhomogeneous flow

in metallic glasses. Acta Metallurgica, 1977. 25: p. 407-415.

100. Heggen, M., F. Spaepen, and M. Feuerbacher, Creation and annihilation of

free volume during homogeneous flow of a metallic glass. Journal of Applied

Physics, 2005. 97(3): p. 033506.

182

References

101. Spaepen, F., Homogeneous flow of metallic glasses: A free volume

perspective. Scripta materialia, 2006. 54(3): p. 363-367.

102. Spaepen, F., A microscopic mechanism for steady state inhomogeneous flow

in metallic glasses. Acta metallurgica, 1977. 25(4): p. 407-415.

103. Qiao, J. and J. Pelletier, Dynamic universal characteristic of the main (α)

relaxation in bulk metallic glasses. Journal of Alloys and Compounds, 2014.

589: p. 263-270.

104. Qiao, J., et al., High temperature deformation in a lanthanum based bulk

metallic glass showing a pronounced secondary relaxation. Materials

Science and Engineering: A, 2013. 586: p. 57-61.

105. Bletry, M., et al., Free volume model: High-temperature deformation of a Zr-

based bulk metallic glass. Acta materialia, 2006. 54(5): p. 1257-1263.

106. Taub, A. and F. Spaepen, The kinetics of structural relaxation of a metallic

glass. Acta Metallurgica, 1980. 28(12): p. 1781-1788.

107. Argon, A.S., Plastic deformation in metallic glasses. Acta Metallurgica,

1979. 27: p. 47-58.

108. Chen, M., Mechanical behavior of metallic glasses: Microscopic

understanding of strength and ductility. Annu. Rev. Mater. Res., 2008. 38: p.

445-469.

109. Mukherjee, S., et al., Viscosity and specific volume of bulk metallic glass-

forming alloys and their correlation with glass forming ability. Acta

materialia, 2004. 52(12): p. 3689-3695.

110. Wang, W.H. and 汪卫华, Correlation between relaxations and plastic

deformation, and elastic model of flow in metallic glasses and glass-forming

liquids. Journal of Applied Physics, 2011. 110(5): p. 053521. 183

References

111. Dienes, G. and H. Klemm, Theory and application of the parallel plate

plastometer. Journal of Applied Physics, 1946. 17(6): p. 458-471.

112. Hagy, H., Experimental Evaluation of Beam‐Bending Method of Determining

Glass Viscosities in the Range 108 to 1015 Poises. Journal of the American

Ceramic Society, 1963. 46(2): p. 93-97.

113. Vlasák, G., P. Švec, and P. Duhaj, Application of isochronal dilatation

measurements for determination of viscosity of amorphous alloys. Materials

Science and Engineering: A, 2001. 304: p. 472-475.

114. Tan, C., et al., Effects of the non-Newtonian deformation on the

crystallization behavior in a Cu-based bulk metallic glass. Materials Science

and Engineering: A, 2013. 586: p. 53-56.

115. Kawamura, Y., et al., Newtonian and non-Newtonian viscosity of

supercooled liquid in metallic glasses. Materials Science and Engineering: A,

2001. 304: p. 674-678.

116. Bakke, E., R. Busch, and W. Johnson, The viscosity of the Zr46. 75Ti8.

25Cu7. 5Ni10Be27. 5 bulk metallic glass forming alloy in the supercooled

liquid. Applied Physics Letters, 1995. 67(22): p. 3260-3262.

117. Gun, B., K. Laws, and M. Ferry, Viscosity-related properties of Mg 65 Cu 25

Y 10 bulk metallic glass determined by uniaxial tension in the supercooled

liquid region. Journal of Alloys and Compounds, 2010. 496(1): p. 582-588.

118. Demetriou, M.D. and W.L. Johnson, A free-volume model for the rheology of

undercooled Zr 41.25 Ti 13.75 Cu 12.5 Ni 10 Be 22.5 liquid. Scripta

materialia, 2005. 52(9): p. 833-837.

119. Mauro, N., et al., A structural signature of liquid fragility. Nature

communications, 2014. 5. 184

References

120. Bryn Pitt, E., G. Kumar, and J. Schroers, Temperature dependence of the

thermoplastic formability in bulk metallic glasses. Journal of Applied

Physics, 2011. 110(4): p. 043518.

121. Wu, X. and C.E. Xue, Size Effect on the Micro-Extrusion Forming of Zr-

Based Amorphous Alloys. Advanced Materials Research, 2011. 228: p. 586-

591.

122. Schroers, J., et al., Thermoplastic forming of bulk metallic glass—

Applications for MEMS and microstructure fabrication. Materials Science

and Engineering: A, 2007. 449: p. 898-902.

123. Liu, Z. and J. Schroers, General nanomoulding with bulk metallic glasses.

Nanotechnology, 2015. 26(14): p. 145301.

124. Kumar, G., J. Blawzdziewicz, and J. Schroers, Controllable nanoimprinting

of metallic glasses: effect of pressure and interfacial properties.

Nanotechnology, 2013. 24(10): p. 105301.

125. Parkinson, R., Properties and applications of electroless nickel. Nickel

Development Institute, 1997. 37.

126. Yan, J., et al., Precision machining of microstructures on electroless-plated

NiP surface for molding glass components. Journal of Materials Processing

Technology, 2009. 209(10): p. 4802-4808.

127. Dobrovinskaya, E.R., L.A. Lytvynov, and V. Pishchik, Sapphire: material,

manufacturing, applications. 2009: Springer Science & Business Media.

128. Quinn, T.F. and T.F. Quinn, Physical analysis for tribology. 2005:

Cambridge University Press.

185

References

129. Humphry-Baker, S.A., K. Peng, and W.E. Lee, Oxidation resistant tungsten

carbide hardmetals. International Journal of Refractory Metals and Hard

Materials, 2017. 66: p. 135-143.

130. Wei, D., Q. Meng, and D. Jia, Mechanical and tribological properties of hot-

pressed h-BN/Si 3 N 4 ceramic composites. Ceramics international, 2006.

32(5): p. 549-554.

131. Zum Gahr, K.-H., Modeling and microstructural modification of alumina

ceramic for improved tribological properties. Wear, 1996. 200(1-2): p. 215-

224.

132. Mekaru, H., et al., Development of Ni–P-Plated Inconel Alloy Mold for

Imprinting on Pyrex Glass. Japanese Journal of Applied Physics, 2009.

48(6S): p. 06FH06.

133. Johansson, S.A. and G. Wahnström, First-principles derived complexion

diagrams for phase boundaries in doped cemented carbides. Current Opinion

in Solid State and Materials Science, 2016. 20(5): p. 299-307.

134. Cottrell, T.L., The strengths of chemical bonds. 1958: Butterworths.

135. De Boer, F.R., et al., Cohesion in metals. Vol. 1. 1988: North-Holland

Physics Publishing

136. Li, Q.-K. and M. Li, Surface structure and properties of NiZr model metallic

glasses: A molecular dynamics simulation. Journal of Non-Crystalline Solids,

2008. 354(18): p. 2060-2065.

137. Zisman, W., Relation of the equilibrium contact angle to liquid and solid

constitution. in 20 Years of Colloid and Surface Chemistry (eds KJ Mysels.

CM Samour and JH Hillister). Kendal Award Address. 1973, American

Chemical Society. Washington. DC. 186

References

138. Kinloch, A., Adhesion and adhesives: science and technology. 2012:

Springer Science & Business Media.

139. Niessen, A. and F. De Boer, The enthalpy of formation of solid borides,

carbides, nitrides, silicides and phosphides of transition and noble metals.

Journal of the Less Common Metals, 1981. 82: p. 75-80.

140. Puttlitz, K.J. and K.A. Stalter, Handbook of lead-free solder technology for

microelectronic assemblies. 2004: CRC Press.

141. Kepaptsoglou, D.-M., et al., Interfacial Charge Transfer and Chemical

Bonding in a Ni–LaNbO4 Cermet for Proton-Conducting Solid-Oxide Fuel

Cell Anodes. Chemistry of Materials, 2012. 24(21): p. 4152-4159.

142. Bolech, M., et al., The standard molar enthalpies of formation of β-La 2 Si 2

O 7 and β-Ce 2 Si 2 O 7. The Journal of Chemical Thermodynamics, 1996.

28(11): p. 1319-1324.

143. Mompean, F.J., J. Perrone, and M. Illemassène, Chemical thermodynamics of

zirconium. Vol. 8. 2005: Gulf Professional Publishing.

144. Teng, H., Overview of the Development of the Fluoropolymer Industry.

Applied Sciences, 2012. 2(2): p. 496-512.

145. Plunkett, R.J., Tetrafluoroethylene polymers. 1941, Google Patents.

146. O'Hagan, D., Understanding organofluorine chemistry. An introduction to

the C–F bond. Chemical Society Reviews, 2008. 37(2): p. 308-319.

147. Jaszewski, R., et al., The deposition of anti-adhesive ultra-thin teflon-like

films and their interaction with polymers during hot embossing. Applied

surface science, 1999. 143(1): p. 301-308.

148. Gaskell, D.R., Introduction to the Thermodynamics of Materials. 2008: CRC

Press. 187

References

149. Schrader, M.E., Young-dupre revisited. Langmuir, 1995. 11(9): p. 3585-

3589.

150. Evenson, Z., et al., Equilibrium viscosity of Zr–Cu–Ni–Al–Nb bulk metallic

glasses. Scripta Materialia, 2010. 63(6): p. 573-576.

151. Ma, J., et al., Micro mold filling kinetics of metallic glasses in supercooled

liquid state. Journal of Applied Physics, 2013. 113(10): p. 104505.

152. Tsao, S. and F. Spaepen, Structural relaxation of a metallic glass near

equilibrium. Acta Metallurgica, 1985. 33(5): p. 881-889.

153. Slipenyuk, A. and J. Eckert, Correlation between enthalpy change and free

volume reduction during structural relaxation of Zr55Cu30Al10Ni5 metallic

glass. Scripta materialia, 2004. 50(1): p. 39-44.

154. De Hey, P., J. Sietsma, and A. Van Den Beukel, Structural disordering in

amorphous Pd40Ni40P20 induced by high temperature deformation. Acta

Materialia, 1998. 46(16): p. 5873-5882.

155. Happel, J. and H. Brenner, Low Reynolds number hydrodynamics: with

special applications to particulate media. Vol. 1. 2012: Springer Science &

Business Media.

156. Kundu, P. and I. Cohen, Fluid Mechanics, 872 pp. Academic, San Diego,

2008.

157. Hasan, M., J. Schroers, and G. Kumar, Functionalization of metallic glasses

through hierarchical patterning. Nano letters, 2015. 15(2): p. 963-968.

158. Zhang, X., et al., An experimental investigation and modeling of micro array

replication with Zr-based bulk metallic glass using a hot embossing process.

International Journal of Machine Tools and Manufacture, 2017. 117: p. 11-

22. 188

References

159. Gu, J., et al., Effects of annealing on the hardness and elastic modulus of a

Cu 36 Zr 48 Al 8 Ag 8 bulk metallic glass. Materials & Design, 2013. 47: p.

706-710.

160. Liu, W.D., H.H. Ruan, and L.C. Zhang, Revealing Structural Relaxation of

Optical Glass Through the Temperature Dependence of Young's Modulus.

Journal of the American Ceramic Society, 2014. 97(11): p. 3475-3482.

161. Liu, W.D., H.H. Ruan, and L.C. Zhang, Atomic rearrangements in metallic

glass: Their nucleation and self-organization. Acta Materialia, 2013. 61(16):

p. 6050-6060.

162. Liu, W.D., H.H. Ruan, and L.C. Zhang, On the plasticity event in metallic

glass. Philosophical Magazine Letters, 2013. 93(3): p. 158-165.

163. Liu, W.D., H.H. Ruan, and L.C. Zhang, Understanding the brittleness of

metallic glasses through dynamic clusters. Journal of Materials Research,

2014. 29(4): p. 561-568.

164. Zhang, H., et al., Investigation of shear band evolution in amorphous alloys

beneath a Vickers indentation. Acta materialia, 2005. 53(14): p. 3849-3859.

165. Zhang, H., et al., Evaluation of hardness–yield strength relationships for bulk

metallic glasses. Philosophical magazine letters, 2006. 86(5): p. 333-345.

166. Keryvin, V., V. Hoang, and J. Shen, Hardness, toughness, brittleness and

cracking systems in an iron-based bulk metallic glass by indentation.

Intermetallics, 2009. 17(4): p. 211-217.

167. Jana, S., et al., Deformation morphology underneath the Vickers indent in a

Zr-based bulk metallic glass. Intermetallics, 2004. 12(10): p. 1097-1102.

189

References

168. Xie, S. and E.P. George, Hardness and shear band evolution in bulk metallic

glasses after plastic deformation and annealing. Acta Materialia, 2008.

56(18): p. 5202-5213.

169. Liu, W.D. and K.X. Liu, Mechanical behavior of a Zr-based metallic glass at

elevated temperature under high strain rate. Journal of Applied Physics,

2010. 108(3).

170. Liu, W.D., et al., The failure stress of bulk metallic glasses under very high

strain rate. Journal of Materials Research, 2010. 25(7): p. 1230-1234.

171. Zhang, M., L.H. Dai, and L. Liu, Mechanical annealing in the flow of

supercooled metallic liquid. Journal of Applied Physics, 2014. 116(5): p.

053522.

172. Puthucode, A., et al., Incipient plasticity and shear band formation in bulk

metallic glass studied using indentation. Metallurgical and Materials

Transactions A, 2008. 39(7): p. 1552-1559.

173. Mukai, T., et al., Dynamic response of a Pd 40 Ni 40 P 20 bulk metallic glass

in tension. Scripta Materialia, 2002. 46(1): p. 43-47.

190